Citation
Cognitive-behavior modification

Material Information

Title:
Cognitive-behavior modification the use of self-instruction strategies by first graders on academic tasks
Creator:
Sparks, Christopher Williams, 1946-
Publication Date:
Language:
English
Physical Description:
viii, 89 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Child psychology ( jstor )
Classrooms ( jstor )
Cognitive models ( jstor )
Control groups ( jstor )
Educational research ( jstor )
Learning ( jstor )
Learning disabilities ( jstor )
Special education ( jstor )
Students ( jstor )
Teachers ( jstor )
Behavior modification ( lcsh )
Cognition in children ( lcsh )
Dissertations, Academic -- Foundations of Education -- UF
Foundations, of Education thesis Ph. D
Self-culture ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1986.
Bibliography:
Bibliography: leaves 85-87.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Christopher Williams Sparks.

Record Information

Source Institution:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029491442 ( ALEPH )
14510379 ( OCLC )

Downloads

This item has the following downloads:


Full Text












COGNITIVE-BEHAVIOR MODIFICATION: THE USE OF
SELF-INSTRUCTION STRATEGIES BY FIRST
GRADERS ON ACADEMIC TASKS





By


CHRISTOPHER WILLIAMS SPARKS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY











UNIVERSITY OF FLORIDA


1986






























Copyright 1986

by

Christopher Williams Sparks






























To Judd, Sara, and Steve

with

thanks and love















ACKNOWLEDGEMENTS


For both the product and the process of my time as a doctoral student, I offer sincere thanks to

My committee, Gordon Greenwood, John Newell, Linda Crocker, and Suzanne Krogh, whose guidance and example have introduced me to worlds of information and new parts of myself;

My teachers, whose gifts of challenge and knowledge have been invaluable;

My family, for signs of pride and words of encouragement when tasks seemed overwhelming; and

My good, good friends, for the big favors, little presents, and constant faith they extended over what, amazingly, turned out to be years.














TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS ....... ................... iv

ABSTRACT ......... ....................... vii

CHAPTERS

I INTRODUCTION ...... ................. 1

Statement of the Problem ...... ........... 1
Self-Instruction Training ...... ........... 2
Self-Instruction Research ...... ........... 4
The Purpose of the Study ...... ........... 6
Significance of the Study ..... ........... 10
Limitations of the Study ... ........... 11
Hypothesis ....... .................. 11

II REVIEW OF THE LITERATURE ... ........... 13

Historical Development .... ........... . 13
The Shift from Operant Conditioning . . . 13
The Impact of Language Development
Studies ..... ................ 15
Influence of Verbal Mediation and Task
Analysis ..... ................ 16
Relevant Metacognition Research .......... . 17
The Formal Self-Instruction Model ....... ... 19
General Studies in Self-Instruction ...... . 20
Meichenbaum's Original Study ....... ... 20 Self-Instruction Studies Since 1971 . . . 21
Academic Studies in Self-Instruction ..... . 23
Guidelines for Effective Research in
Self-Instruction .... ............. . 28
Summary of the Literature Review .......... 29

III METHODOLOGY ....... .................. 31

Research Questions and Hypothesis ....... ... 31 Method of the Study ..... ............. . 32
Subjects ....... .......... ......32
Instruments ..... .............. 33
Procedure ...... ............... 43
Statistical Design .... ............ 52









Pagte

IV RESULTS AND CONCLUSIONS ... ............ 54

V SUMMARY AND RECOMMENDATIONS ... .......... .. 62

Summary ........ .................... 62
Conclusions ....... .................. 64
Implications ...... ................. 65
Recommendations ...... ................ 67

APPENDICES

A INVITATION TO PARTICIPATE ... ........... . 69

B SELF-INSTRUCTION TRAINING ... ........... . 71

C GUIDELINES FOR EFFECTIVE SELF-INSTRUCTION. . . 74 D MEICHENBAUM'S SELF-INSTRUCTION MODEL ..... . 76 E INSTRUCTIONS FOR THE EXPERIMENTAL GROUP .... 77 F INSTRUCTIONS FOR THE CONTROL GROUP ...... .. 82 G RESEARCH DESIGN ...... ................ 84

REFERENCES ......... ...................... 85

BIOGRAPHICAL SKETCH ....... ................. 88












Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy


COGNITIVE-BEHAVIOR MODIFICATION: THE USE OF
SELF-INSTRUCTION STRATEGIES BY FIRST
GRADERS ON ACADEMIC TASKS

by

Christopher Williams Sparks May 1986

Chairman: Gordon E. Greenwood Major Department: Foundations of Education

The purpose of this study was to apply the cognitivebehavior modification self-instruction model, which involves the use of cognitive modeling and children's private speech to develop children's self-guiding statements and images in task performance, to actual classroom practice and content.

Fifty-six first grade students were randomly assigned within intact reading groups to two treatment conditions. Students in the experimental group were trained by their teachers in self-instruction on non-reading tasks during four sessions separate from reading instruction. Both experimental and control groups participated in regular, basal-directed reading instruction. The dependent measure for this study was children's performance on six selected independent seatwork tasks provided by the reading series.


vii








A multivariate analysis of variance (MANOVA) using

a two-way nested design was used to test the hypothesis that there were no significant differences between experimental and control groups on their overall task performance. Univariate analyses of variance (ANOVAs) were then used to determine if the overall difference between groups existed for each task. The MANOVA analysis yielded a highly significant difference between groups, followed by the ANOVAs' smaller but significant differences on five of the six tasks.

From the above results, the following conclusions were drawn. First, children who used self-instruction applied academic skills with significantly higher accuracy through increased skill in problem definition, strategy selection, self-evaluation, self-reinforcement, and error coping. Second, self-instruction, as a teaching-thinking process which focuses on the processes students use in task performance, is generalizable across at least some tasks. Third, self-instruction can be effectively used to teach planful behavior in academics with groups of normal children in classroom settings. Fourth, teachers can be easily and effectively trained to incorporate self-instruction into their curricula. Fifth, cognitive strategy changes can be demonstrated on ordinary performance in naturalistic settings. All conclusions point to promising gains through further applied cognitive-behavior modification research.


viii















CHAPTER I
INTRODUCTION



Statement of the Problem


One of the major challenges in education today is the development of students' independent thinking skills and work habits. Children often acquire information in a passive, rote fashion, and then are unable to respond to questions and solve problems unless those questions and problems closely resemble the learning task. This sort of "patterned learning" effect has been particularly apparent to this researcher (a teacher) in school-based reviews of standardized test errors and math problemsolving mistakes, in children's composition efforts, and, in the middle school years, their inability to work on class or home assignments without guidance or supervision. One approach to increasing children's personal control over learned information is to train them in selfinstruction, a cognitive control technique developed within the cognitive-behavior modification paradigm by Donald Meichenbaum and others in the early 1970s.

By the end of the 1950s operant conditioning had begun to evolve from various forms of laboratory


-1-





-2-


stimulus-response investigations to behavior therapy programs with children. Educators in the 1960s became familiar with behavioral objectives, behavior modification, token economies, and precision teaching. "Modeling," "shaping," and "positive reinforcement" became part of everyday professional language and the theme of teacher inservice programs in the early 1970s. Further, during the 1970s behavior therapy underwent a "shift from an emphasis on modification of attentive and disruptive motor behaviors to a concern with educational tasks that involve cognitive or thinking skills" (Meyers & Craighead, 1984, p. 4). This transition was fostered by the work of social learning theorists which provided cognitive explanations of modeling effects, and by the recognition that behavior therapy, or simple operant conditioning, did not produce generalizable or long-lasting behavior change. Selfinstruction evolved from cognitive-behavior modification investigations of cognitive processes and strategies children use in task performance.



Self-Instruction Training


In self-instruction training, cognitive modeling and private speech are combined as an approach to teaching thinking skills. Children are taught how to use selfstatements and images to think and plan behavior; they are









not instructed in what to think or encouraged to focus on "right answers." Although right answers would appear to be a logically correct educational goal, both metacognitive and self-instruction researchers reject that aim and set the development of underlying thinking skills or executive processes as a priority of teaching efforts.

Self-instruction in task performance requires that the student engage in problem definition, appropriate strategy selection, self-evaluation of performance, self-reinforcement, and coping behaviors. To date, the self-instruction process has been used with young children on social, interactive, motor, and academic tasks. Meichenbaum's model includes the following steps:

1. cognitive modeling (an adult performs the
task and talks to self);

2. child performs the task with model's direction for guidance (overt external guidance);

3. child performs task and talks aloud to self
(overt self-guidance);

4. child whispers and performs task (faded,
overt self-guidance);

5. child performs task via inaudible or private
speech or nonverbal self-instruction (covert
self-instruction). (Meichenbaum, 1977,
p. 32)

At step three the child's skills in problem definition, planful behavior, and implementation of strategies can be observed.





-4-


Self-Instruction Research


The cognitive-behavior modification investigations

utilizing self-instruction during the 1970s and early 1980s have demonstrated success in modifying subjects' performance. Research designs have included, usually, single subjects of specialized (hyperactive, learning disabled) populations in laboratory settings. Tasks presented have been sensorimotor or interpersonal, and treatment effects have been measured on tests other than direct performance (such as Porteus Mazes, standardized tests, or the Matching Familiar Figures Test). These designs have been apropos for theoretical development and, in effect, piloting the use of a new intervention. Such intervention research is normally prompted by some need in a clinical population and focuses on resolving various dysfunctions for a long period before being used in a preventative or developmental fashion.

If self-instruction is to be proved a useful way to

enhance the development of thinking skills in normal early learners, previous research designs must be altered. To begin with, utility with groups of normal subjects on academic tasks in naturalistic settings must be demonstrated. Once such success is established, Craighead, Meyers, Craighead, and McHale (1983) indicate that client assessment, generalization of treatment effects, and loci of





-5-


interventions, and use of paraprofessionals will become the issues of the 1980s. In other words, the testing of the self-instruction model as a teaching-thinking program in classroom settings can begin to address the question of its practicality. Meichenbaum sets the stage for selfinstruction research in academics by stating,

We can conceive of academic tasks where the teacher provided the children with a set of
tasks . . . and the child's job was to identify what the problem is, how he or she will go about solving the task, where the likely
pitfalls are, etc. . . . Teachers could give assignments and ask the children to describe in detail how they are going to go about performing the assignment. . .. Discussion could center on the process, not only the product, of the assignment. (Meichenbaum,
1985, p. 421)

A number of other studies have investigated selfinstruction in designs which would not be replicable or appropriate in regular classrooms (see Meichenbaum, 1985, for a review of these studies). There have been only two doctoral studies done (Rhodes, 1979; Sullivan, 1981) which have begun to approach the full applied Meichenbaum model. In one of these studies (Sullivan, 1981), "selfinstruction" was interpreted as verbalization of a given problem in mathematics rather than as the development of self-guiding strategy statements and was used in conjunction with a mathematics problem-solving strategy, yielding no information about the metacognitive advantages of self-instruction per se. In the other study (Rhodes,





-6-


1979), the self-instruction process was abbreviated until only its cognitive-modeling component was taught, and, again, the effects of the full self-instruction model were not assessed. The Meichenbaum model does involve cognitive modeling and overt-to-covert skills practice, but these procedures must further include opportunity for the child to engage in problem definition, strategy selection and implementation, and error management. This study provides necessary information regarding the feasibility of utilizing the full self-instruction model in classroom situations, with normal children, on academic tasks.



The Purpose of the Study


The purpose of this study is to apply the cognitivebehavior modification self-instruction model as developed by Meichenbaum to actual classroom practice and content. The design includes the use of the full self-instruction model without abbreviation, extension, or collaboration with other problem-solving methods and is in direct response to the following needs statements in general reviews of the field.

1. Kendall (1977), in a discussion of obtaining generalization of learning through use of cognitivebehavior modification techniques, stated "the focus of the training materials and the setting contribute





-7-


meaningfully to the type of generalization achieved" (p. 331), and suggested that the way to promote "the natural development of verbal mediation" (p. 331) and to "foster each child's personal cognitive control" (p. 331) in the classroom is to use psychoeducational materials in a teacher-student situation.

2. Meichenbaum and Asarnow (1979) maintained that self-instruction training should teach "cognitive skills or executive routines that are transsituational" (p. 28) and that such training should "focus directly and explicitly on the skills and tasks that are to be learned, not on some presumed underlying deficit" (p. 30).

3. Hobbs, Moguin, Tyroler, and Lahey (1980) questioned the demonstrated utility of cognitive-behavior modification with children, indicating that the common problems with this research area were the specificity of the independent variable, lack of results from natural settings and normal performance, and lack of information on the impact of treatment in classroom or home situations.

4. Meichenbaum and Burland (1981), commenting on the self-instruction format, stated, "At this point we can share the general training strategy that is being employed and make a call for more research to assess the pedagogical potential of cognitive-behavior modification procedures" (p. 112).








5. Kazdin (1982) emphasized the need to show that
"changes in cognitive skills are reflected in measures of ordinary performance in naturalistic situations" (p. 77 ) and stated that when the target population is "not identified on the basis of deficits of dysfunction, the criteria for significant change may be demonstration that the intervention produces a significant increment above the previously accepted normative level" (p. 78 ).

6. Meichenbaum (1985) pointedly stated the need for classroom teacher participation in self-instruction research and set guidelines for the development of teachingthinking classes to be integrated into the school curricula.

In this study, first grade teachers from among volunteers in a local school district were randomly assigned to treatment or control groups; self-instruction supplemented regular reading instruction for the treatment groups. Teachers in the treatment conditions were trained in the principles and administration of self-instruction, and they trained their students in self-instruction. The investigation took place in normal first grade classroom settings; academic tasks from the established curriculum were presented to the students in both treatment and control groups to be done independently.

For the purpose of this study, the following terms are defined by specific characteristics.





-9-


"Normal first graders" are those children in regular classrooms who have not been retained in the grade and who are not staffed into any special program. It is assumed that students not identified or referred for special education programs are "normal" or "average" in that they possess a range of abilities exclusive of the very high and the very low.

A "classroom setting" refers to a nonlaboratory educational situation consisting of one teacher and approximately 25 students. The students involved in the study were not participating in any other school-based programs (tutoring, precision teaching, etc.).

"Academic work" means written basal reading assignments in the district-adopted reading series, Ginn. These workbook pages provide a set of work for the children which has content validity and direct relevance to skills taught. Children are tested for correct placement in this series.

"Self-instruction training" is the use of images and self-guiding statements to direct behavior.

"Private speech" refers to egocentric or self-addressed verbalizations; it is not intended for a listener to hear or to communicate with others.





-10-


Significance of the Study


The original self-instruction studies targeted

hyperactive/impulsive children in an effort to teach those children to engage in planful behavior, to be aware of and use cognitive strategies as skills increased in selfcontrol or interpersonal behavior. There is still interest in changing the reflective/impulsive balance or cognitive tempo of such specialized populations, but applied research in self-instruction is needed to extend knowledge of its effectiveness on academic tasks and its use as a preventative or developmental approach to problems apparent in specialized populations.

In normal populations, using self-instruction to

teach students planful behavior in academics remains to be investigated and may provide a remedy for those problems which are easily visible in school settings: students' inadequate study skills, inability to work independently, and inability to solve problems in any fashion other than that patterned from the teacher. The positive results of this study will substantially extend the limited existing information in this area and will contribute to the literature on the explicit training of metacognitive strategies.





-11-


Limitations of the Study


The following limitations must be considered in this study.

1. Teachers were selected from a volunteer pool

recruited by open invitation across the district. These teachers' professional curiosity and willingness to participate in experimentation may be indicative of other qualities which make them "more effective" or "more dynamic" than others.

2. The treatment condition could have been misinterpreted or incorrectly administered by individual teachers in spite of careful training; the control groups could have been affected by their teachers' needs for them to do well.



Hypothesis


Treatment teachers participating in this study supplemented regular basal-directed reading instruction with students' training in self-instruction. Measures of normal classroom performance were taken on six sets of basal reading tasks completed independently by the children. The control groups received regular basal-directed reading instruction and completed the same assignments as the treatment groups, but without self-instruction training.





-12-


The supposition was that children with self-instruction training would apply academic skills with significantly higher accuracy because of increased skills in problem definition, strategy selection and implementation, and error management. The multivariate null hypothesis developed was as follows:

There will be no significant differences between the experimental and control groups in terms of performance on a set of six separate academic tasks including: word identification in context and reality/fantasy discrimination; decoding verbs with inflected s; word insertions in sentences and puzzles; sequencing pictures or sentences; completing sentences with inflected-s verbs; advanced word insertions in sentences and in puzzles.















CHAPTER II
REVIEW OF THE LITERATURE



Historical Development


The Shift from Operant Conditioning


Cognitive-behavior modification experimentation began in an attempt to advance results gained in operant conditioning: That is, the intent was to develop the subject's capacity to retain skills or patterns learned and to generalize those skills and patterns across different situations and tasks. Such retention and transfer were not generally a product of operant conditioning.

Craighead (1982) identifies three major factors leading to the shift, in the 1970s, from operant to cognitive-behavioral interventions with children. The first was a cognitive information processing explanation of modeling effects; the second was development of selfcontrol interventions; and the third, independent of the first two, was developments in cognitive therapy.

Cognitive information processing. In the area of

modeling or observational learning, Craighead (1982) cites Bandura's (1969) work on the roles of attention and


-13-





-14-


retention as the beginning of new interpretations of behavior therapy procedures. Bandura's social learning theory proposed that cognitive and environmental influences are mediated by the cognitive processes of perception and attribution, and that the person and the environment influence one another. From the tenets of social learning theory sprang laboratory-based investigations of children's self-mediational strategies in enhancing self-control on such tasks as delay of gratification and resistance to temptation (see Meichenbaum, 1985, for a review of these studies).

Self-control. Self-control studies, which had been placed within the realm of operant conditioning for the duration of behaviorism's influential decade, began, in the 1970s, to be conceptualized as more cognitive in nature. Craighead (1982) cites Kanfer's work dividing self-control into the components of self-monitoring, self-evaluation, and self-reinforcement as advances in clinical application and research in self-control interventions. These "advances" later contributed significantly to the principles involved in Meichenbaum's selfinstructional format.

Cognitive therapy. Cognitive-behavior therapy with

adults began in attempts to "combine the clinical concerns of cognitive-semantic therapists with the behavior therapy technology" (Meichenbaum & Burland, 1979, p. 426).





-15-


Craighead (1982) maintains that fundamental to cognitive therapy is the assumption that maladaptive cognitions such as inappropriate, irrational and illogical self-statements (based on personal assumptions and beliefs) produce psychological disorders which are best alleviated by modification of those cognitions. Cognitive-behavior therapy with children was given impetus as this relationship between cognitive and behavioral/physiological events began to be established and continued to develop as a therapeutic alternative to psychoanalysis or behavior therapy in the treatment of childhood disorders. Meichenbaum's use of self-statements in behavior change evolved independently of this work in cognitive-behavior therapy, but against its favorable backdrop.



The Impact of Language Development Studies

Soviet research. Added to the converging influences of social learning theory, self-control interventions, and cognitive-behavior therapy was the work of Soviet psychologists in language development. According to Vygotsky (1962), children's speech progresses from an obvious externally oriented, overt stage, through an egocentric, self-addressive phase, to inner speech. Vygotsky suggests that inner speech is a foundation for children's thinking. The overt-to-covert





-16-


language development model suggested to Meichenbaum that "private speech is initially facilitative and then drops out of the repertoire with the development of task proficiency" (Meichenbaum, 1977, p. 23).

Meichenbaum's use of private speech. In 1969 Meichenbaum's doctoral research involved training schizophrenics to emit "healthy talk" to control their verbalizations and behavior; his success in this area led him to ask if other clinical populations could be explicitly trained to self-instruct and spontaneously generate statements to guide behavior (Meichenbaum, 1977).



Influence of Verbal Mediation and Task Analysis


Verbal mediation. Coincidental to the developments

in cognitive-behavior modification were efforts being made in the analysis of "verbal mediation deficiencies." In the area of verbal mediation, the use of task-appropriate mediators is seen as involving comprehension of task, production of strategies, and, importantly, implementation of those strategies (Meichenbaum & Asarnow, 1979). While the cognitive-behavioral self-instruction process does not call for such emphasis on a phase-division of mediational intervention with children, it does call for a "cognitivefunctional analysis of children's task performance" (Meichenbaum & Asarnow, 1979, p. 12).





-17-


Task analysis. The cognitive-functional analysis of task performance requires an awareness of the child's self-statements and images (which can be observed during the overt self-guidance step of self-instruction) and an understanding of the task components. Meichenbaum used the work of Gagne (1964) as a model for identifying the hierarchy of behaviors within a task, but translated each step into self-statements and strategies that could be modeled by the trainer and rehearsed by the child (Meichenbaum & Asarnow, 1979).



Relevant Metacognition Research


The field of cognitive-behavior modification progressed parallel to and convergent with the field of metacognition. Each is concerned with cognitive strategy deficits leading to poor performance on tasks, and each is concerned with the learner's ability to identify the problem at hand, select and implement strategies, and evaluate performance. Cognition refers directly to such acts of thinking as attending, remembering, and choosing; metacognition refers to one's personal awareness of those acts.

Researchers laboring under the rubric of "metacognition" tend to demonstrate interest in identifying the cognitive processes leading to deficit performance;





-18-


researchers working in cognitive-behavior modification tend to maintain concern for the subject's performance when applying appropriate cognitive strategies. Each group is concerned with effective use of cognitive strategies and the subject's awareness of personal control over those strategies. Those of the metacognitive ilk focus more on breaking the cognitive processes into components, while the cognitive-behavior modifiers focus more on breaking the task into components.

Work on metacognitive development suggests that
"children fail to consider their behavior against sensible criteria, they follow instructions blindly, and they are deficient in self-questioning skills that would enable them to determine these inadequacies" (Brown, 1980, p. 457). Self-instruction (saying guiding statements to oneself) is one technique from the cognitive-behavior modification camp which is "often used to teach a person to monitor his progress, compare what he/she is going to what he/she should be doing, and self-reinforce" (Craighead et al., 1978, p. 164); in other words, self-instruction provides a mechanism for one to engage in metacognitive acts before, during, and after task performance.

In a review of studies on the use of cognitive

instruction, Meichenbaum (1985) points out that children who receive cognitive instruction on specific skills without attention or training given to the development of




-19-


superordinate processes are not able to generalize learned skills to other tasks or maintain improvements across time. On the other hand, children taught to define problems, engage in planful behavior, and monitor that behavior achieve substantial transfer of learning. The selfinstruction sequence fosters the development of those

planful behaviors or thinking skills.


The Formal Self-Instruction Model

The original study. The research in language development and his own successful clinical use of private speech implied to Meichenbaum that a therapeutic package could be developed which would utilize self-instructional statements in behavior change; with Goodman, (Meichenbaum & Goodman, 1971), Meichenbaum formalized a self-instruction model for use with children to control impulsivity and motor behavior and to develop self-control. The final structure of the model reflected current progress in social learning theory, self-control interventions, cognitive therapy, language development, verbal mediation, and task analysis.

Current reviews. After reviewing the results of almost a decade of self-instruction investigations, Meichenbaum and Asarnow (1979) theorized that the overt speech in self-instruction training organizes the





-20-


information available, facilitates problem definition, and aids in strategy selection and implementation. Further, they suggested that orientation and attention to task may be enhanced, and that coping skills and self-reinforcement skills become a useful part of the child's repertoire.

The self-instruction format provides a content-free question set which the child can utilize across a variety of tasks. The goal of cognitive-behavior self-instruction is to teach children to generate, choose, and apply cognitive strategies, not to increase a specific skillsbase. Meichenbaum's (1985) review of the field indicates that the time has come to investigate the use of selfinstruction in classrooms on academic tasks.



General Studies in Self-Instruction


Meichenbaum's Original Study


Research in cognitive-behavior modification utilizing self-instruction began with Meichenbaum and Goodman's (1971) study in which hyperactive, impulsive preschoolers were trained to talk to themselves as they did a variety of sensorimotor (copying patterns, coloring figures) and problem-solving (completing pictorial series, solving mazes) tasks. Results of this study indicated that hyperactive, impulsive children could be taught to think before they acted and not be subject to the dominant motor





-21-


response (Meichenbaum & Goodman, 1971). The children trained in self-instruction had significantly better performance on a variety of standard psychological tests and demonstrated increased reflectivity on the Matching Familiar Figures Test. Further, more than half of the trained impulsive children continued to use self-instruction strategies on the posttest and in later sessions (Meichenbaum & Burland, 1981).

The Meichenbaum and Goodman study used (a) single

subjects of (b) a special population in (c) a laboratory setting; (d) the tasks involved were sensorimotor and (e) were not tested on a direct performance measure, but on tests quite removed from the task. Most of the relevant investigations done in the decade following this 1971 study were different in theme but almost identical in composition; these studies are presented here to show, for historical purposes, the types of studies done in selfinstruction.


Self-Instruction Studies Since 1971

The most comprehensive review of studies in selfcontrol is Meichenbaum's (1985) "Teaching Thinking: A Cognitive-Behavioral Perspective," a chapter in Chipman, Segal and Glaser's Thinking and Learning Skills (Volume 2): Research and Open Questions which was sent to the





-22-


investigator by Meichenbaum when a request for material relevant to applying self-instruction to academics was made. This review includes most of the studies specified below and represents the main body of research in selfinstruction; most investigations cross-reference others which pre-date them, and they represent the range of topics studied before an interest in academic application was developed.

In a review of cognitive-behavior modification with

children, Meichenbaum and Burland (1981) cite its successful use to establish "inner speech control" over the disruptive behavior of hyperactive children and aggressive children; with cheating behavior; with Maze performance of hyperactive boys; over the conceptual tempo of emotionally disturbed children; and over the conceptual tempo of normal children.

Other investigations cited by Meichenbaum and Burland combined self-instruction training with imagery practice and with operant procedures. Also cited are work with learning disability children; work with hyperactive children; work with retarded children; and articles on social reasoning and problem-solving using self-instruction.

In another review paper, Meichenbaum and Asarnow (1979) reference work with hyperactive children on arithmetic tasks, work on verbal control tasks, and use of self-instruction on kindergarteners' handwriting skills.





-23-


In an article by Hobbs, Moguin, Tyroler, and Lahey (1980), work with hyperactive children's conceptual tempo is cited. Again, the interest of this study is in the use of selfinstruction with normal children, in nonlaboratory settings, on academic work normally distributed to the children; the preceding references have been presented as sources for the interested reader and to illustrate the nature of investigative interest to this point.

Academic Studies in Self-Instruction

An extensive review of the literature including the Current Index to Journals in Education (CIJE), Resources in Education (RIE), Educational Resources Information Center (ERIC), and the Dissertation Abstracts International yielded two studies concerning academic tasks and normal children in conjunction with self-instruction. These two studies are the most directly relevant done to date since they involved academic research in self-instruction with regular classroom populations. They are described below.

Sullivan, 1981. At the University of Oregon, Sullivan (1981) did a doctoral study called "A Comparison Between Attack Strategy Training and Attack Strategy Training in Combination with Self-Instruction in Teaching Academic Tasks to First Graders." Although Sullivan's investigation





-24-


was directed at comparing results for impulsive and reflective children, and his results were measured on a researcher-developed math criterion test, various standard intelligence tests and achievement tests, all of which would have negated interest by this researcher, there were other disengaging differences illustrative of the ways in which self-instruction research is interpreted.

Sullivan used direct instruction to teach math computation skills to one group, and, for the second group, coupled that instruction with self-instruction in the sense that children were "verbalizing key phrases aloud as they worked through the task" (Sullivan, 1981, p. 44). Sullivan's math problems were "academically relevant materials" (p. 8), but were not a subsection of the students' normal math curriculum. Sullivan's study "did not assess the effectiveness of verbal self-instruction as a treatment without specific attack strategies" (p. 84). In other words, children verbalized statements of particular strategies which were taught to them; they did not generate their own self-statements and strategies. This approach fits more closely with studies concerned with the influence of self-verbalization on academic achievement (Grimm, Bijou, & Parsons, 1973; Lovitt & Curtis, 1968); these studies found that simple verbalization of problems in the math area enhanced children's performance. The true selfinstruction procedure is content-free: It calls for





-25-


problem definition by the student, selection of strategies by the student, and self-monitored implementation of those strategies. Sullivan's "self-instruction" group did outperform his attack-strategy-only group, and his study was based on Meichenbaum's work, but his interpretation of the self-instruction purpose and format and its application are not in alignment with that of this researcher or with the stated guidelines for research by Meichenbaum (see next section).

Rhodes, 1979. The University of Louisville dissertation by Rhodes (1979) is problematic in some of the same ways as Sullivan's. The author's intent was to demonstrate that Meichenbaum's self-instruction could be used to direct readers' attention to task and to teach students to generate questions regarding the purpose and meaning of assigned reading passages. Although the results of this dissertation indicate that behavioral methods do increase reading comprehension (on standardized tests after five weeks of intervention), the Meichenbaum model was substantially abbreviated, and the task so narrow that it was not clear that cognitive self-instruction was responsible for the increase in comprehension, or that, as a modus operandi, the children learned a taskgeneralizable self-instruction model; they may have learned simply to question the title and paragraph of any story.





-26-


In the Rhodes study the "self-instruction" format involved, first, the teacher's modeling or scanning the assigned story, reading the title and first paragraph, and questioning for purpose and meaning as cued by the teacher's manual; this took up one week. For the second week, students individually scanned the material, generated questions regarding purpose and meaning, and read silently. The teacher selected certain of the student-generated questions for discussion at the conclusion of reading. During the third, fourth, and last weeks these procedures were repeated, but the children's spoken questions faded to silence.

In the Rhodes study children had to generate questions regarding story purpose or meaning from the title and first paragraph; there was no problem definition or strategy selection in that task, nor was there any of the necessary self-reinforcement or error management modeling. Further, it is not clear whether the control group was directed to scan and then re-read the story (without the modeling and question-generating time), in which case the number of times each story was read and the following question-discussion could have increased comprehension for the experimental group.

It appears that the Rhodes study is one in which the cognitive modeling portion of the full self-instruction model was used as a teaching strategy for a particular





-27-


reading task;the procedures used are, actually, more akin to the SQ3R (Survey, Question, Read, Recite, and Review) method cited in Pressley and Levin's (1983) Cognitive Strategy Research: Educational Applications and in much of the reading literature. In this process, the reader scans titles, subheadings, and summaries to get an idea about the story before reading and then restates any boldface headings in question form. Again, as with Sullivan's study, in the Rhodes study there is both a distortion of the Meichenbaum model and an overlap with research in the content area, or, at best, use of a simple faded cognitive-modeling teaching strategy.

Summary of academic studies. The two dissertations reviewed above present varying points of view as to the uses and interpretations of self-instruction. They do not, however, clearly apply the Meichenbaum selfinstruction format to academic instruction. Each of these authors seemed to perceive self-instruction as a new mode of instruction for the purpose of getting more "right answers" in a particular subject area rather than as an approach to teaching thinking skills.

In order to further explore and understand the use of self-instruction with young children, the research must focus on development of strategies which are contentfree and not patterned as a "set" of attack plans for a particular "set" of problems. The dissertations reviewed





-28-


do not address the research need or adhere to the guidelines for self-instruction research.



Guidelines for Effective Research in Self-Instruction


Meichenbaum (1985) makes recommendations in two areas for those who wish to do research in self-instruction training: First, he describes a point of view to which the teacher must be committed in order to effectively transmit the training, and, second, gives guidelines for the training to foster generalization.

Ideally, the teacher should model a metacognitive

perspective, one which indicates openness to thinking and anticipation of mistakes which, when they occur, are viewed as problems to be solved. The instructor should be careful of attributions made when failures occur, with statements directed at nurturing the problem-solving attitude. Lastly, the teacher must conduct a task analysis of skills to be taught, as self-instructions are comprised of superordinate cognitive strategies which are implemented to successfully apply content skills.

To maximize generalization, training in self-instruction must be on a task which has the same requisites or elements of the performance task., but is overtly different from that performance task. The pupil must be a collaborator in generating the strategies to be used, and the private





-29-


speech trained must be developed with and compatible with the child's natural style. The cognitive strategy training should take place after component skills and experience with the task are in the child's repertoire, and the pupil must be helped to recognize the new task as one facilitated by self-instruction and requiring transfer of those strategies. The instructor needs to ensure against rote repetition of patterned strategies, encouraging personal involvement by the student through use of faded cues and acceptance of the child's wording. The teacher should directly encourage the child to generalize strategies to various tasks or situations, perhaps by engaging in discussion of other suitable uses. Lastly, a sense of selfsatisfaction should be nurtured in the child, with attention given to developing the child's self-reinforcing and coping skills and flexibility in the use of strategies is fostered.



Summary of the Literature Review


From the foregoing review of the literature the following conclusions are apparent.

1. Self-instruction provides a mechanism for one

to engage in metacognitive acts before, during, and after task performance; in other words, self-instruction can be used to teach "planful" behavior.





-30-


2. Since the original Meichenbaum study dealing

with hyperactive/impulsive children on sensorimotor tasks, many investigations using self-instruction with specialized populations on sensorimotor or interpersonal tasks have been done.

3. Most self-instruction studies have involved single subjects, special populations, and laboratory settings. Most tasks studied have been sensorimotor and have not been tested through measures of normal performance.

4. The two studies done on the use of self-instruction in academics have not interpreted self-instruction in a metacognitive sense and have not met the current needs for applied research.

5. No studies have been done utilizing selfinstruction with normal children, in a classroom setting, on normal performance of academic work.















CHAPTER III
METHODOLOGY



Research Question and Hypothesis

The following research question emerged from the review of research on the use of self-instruction with young children: Can the performance of normal first graders on basal seatwork tasks be improved through the use of self-instruction? Or, in other words, can cognitive strategy changes made be reflected to a measurably significant degree on classwork?

The following null hypothesis was developed to be tested by this study:

There are no significant differences between the

experimental and control groups in terms of performance on any of six separate academic tasks including: word identification in context and reality/fantasy discrimination; decoding verbs with inflected s; word insertions in sentences and in puzzles; sequencing pictures or sentences; completing sentences with inflected-s verbs; advanced word insertions in sentences and in puzzles.


-31-





-32-


Method of the Study


Subjects


First grade students were chosen as the target group

for this study for two reasons: the speech characteristics of their age and their academic skills level.

In the state of Florida first grade students must be six years old on or before September first of the entry year. Between ages five and seven children's speech is more overt than covert; before age five speech is largely reflective and modeled from others and after age seven is likely to be internalized, or covert. The overt-speech characteristic is necessary for observation of the selfinstruction development.

Self-instruction training provides the student or

client with strategies for problem identification, strategy selection, and implementation. The student or client utilizes self-instruction in conjunction with his or her skills in the problem or content area. First graders who would be considered "normal" have an adequate but not sophisticated skills-base in reading and math; such a skills base becomes the point of departure for selfinstruction training. In the reading area assigned level chosen for this study, children do recognize letters and sounds and are aware of what reading is, but have not





-33-


begun to generate their own word attack strategies or to develop sight vocabularies.

Nine teachers with a total of 56 student subjects from

2 rural and 3 city schools were recruited by open invitation to all first graders in the Alachua County, Florida, schools. Each teacher's instructional assignment for the year included a class of approximately 25 students with a range of ability in the reading area. Within each of the nine classes, the group of students in the beginning first grade instructional level was randomly assigned to the treatment or control group for the study. No school contained both treatment and control groups; each treatment condition contained both urban and rural schools.



Instruments


Any academic task for which instruction and standardized seatwork materials are provided would lend itself to self-instruction. For this study reading seatwork tasks were chosen because, in the Ginn district-adopted reading series, such lessons are prescribed in the teachers' manuals, pacing is set for each book, and workbook materials (the "Skillpack" and "Studybook") correlate with the instruction and pacing. These seatwork materials are intended to be presented within the regular classroom setting to be done independently by the children. Further,





-34-


for this series children are tested for correct placement before beginning a level, ensuring that all subjects belonged at the given level of instruction.

The choice of Ginn Level 3, Fish and Not Fish (Clymer, Venezky, Johnson, & Pearson, 1984) was made for several reasons. First, Levels 1 and 2 are primarily letter/sound and experiential materials without basal-structured worksheets involving reading on the students' part. Many first graders enter with sufficient skills to waive Levels

1 and 2, since those levels simply review and reinforce the kindergarten program. Level 3 is an acceptable firstinstruction book for first graders. Levels 4, 5, and 6, the remaining books for instruction in first grade, would also be appropriate for use with self-instruction, but students' placement in those levels in the early part of the year would indicate somewhat sophisticated reading skill. Successful training might be attributed to the learner's earlier readiness for assignment to that level. For Levels 4, 5, or 6 to be tested, the timing of the investigation for normal first graders should be OctoberNovember or later.

Levels 3, 4, and 5 of the Ginn reading series, all being potential choices for the investigation (depending on the implementation dates of the study), were subjected to a task analysis by the investigator. This means that within each level specific tasks were identified for which





-35-


cognitive strategies would be useful. These tasks included phonics exercises, word-identification in context, use of multiple meanings, identifying the main idea, sequencing, and distinguishing reality/fantasy. A set of tasks for each book was assembled from among the tasks which had a large enough item bank (9-27 items) to be useful.

Level 3 thus became the appropriate selection for

self-instruction training because it is normally assigned to children in first grade between ages five and seven, it has a standardized independent-work program, it is an acceptable first book for first graders, and its materials contain a sufficient item bank of specified skills to make measurement feasible; these advantages dovetailed with the timing of the study.

The specific tasks from Level 3 which were involved with this study included phonics exercises, word identification in context, and sequencing, all with different underlying reasoning required of the student. The tasks, which were divided into six sittings, are described below and illustrated by Figures 3-1 through 3-6.

Set 1 is word identification in context and reality/ fantasy discrimination, demonstrated by matching pictures to sentences using lines, within Unit 1 of the book (Figure 3-1).







-36-


Pr Skiifl for L 2. * PIO i..*,



Read each sentence. Then Crow a line to the
picture It matches.



The fish is for the dolahin..
SThe fish is forlM hina2Look al the dahnt
-.Look at the lht




3. -We can get fiuh for the dolphn."
-We Can get fish to eat.-


4. Jj The man con get the fish.

The dolph"n con get the fish.


- 14

Fleaoach lenienc . Then or in Is ite dture i isoiviire.


3..


~>


Am con. Ania can not.



Jim con J. con noL



Ken cah. Ken Can raL


0500 bt.t.t~O0**t*t.*O00


Read each sentence. Then draw a line to tr picture it matches.


wtee does Betn eof? Where does the cat eat?




The co wilt gel water. And wll get water.




Where does Saor like to play? o Where doe te cat like to ploy?




Jim and Ken like the water. __The cat does not ilke the water. /.


Figure 3-1. Performance Tasks, Set 1 (Clymer et al.,1984, used with permission.)


2.nt


1. Sora+ Can+
,...- - Sar Con naL


1 2;





-37-


Set 2 deals with word identification in context demonstrated by matching sentences to pictures using circles, lines, or bubbles, within Unit 2 of the book; and decoding verbs with inflected s (Figure 3-2).

Set 3 involves word insertions using sight vocabulary with key words at the top of the page, within Unit 1, and word insertions in crossword puzzles (Figure 3-3).

Set 4 is sequencing pictures or sentences, within Unit 2, by assigning a number to the pictures or sentences (Figure 3-4).

Set 5 involves word insertions in sentences using

verbs with s inflection, with the verbs written above or below the sentence, within Unit 2 (Figure 3-5).

Set 6 deals with word insertions using sight vocabulary with words at top of the page, within Unit 2, and word insertions in crossword puzzles (Figure 3-6).

The performance score is the total right/wrong measure for each task set, representing the students' abilities to apply academic skills in reading to independent seatwork tasks, with or without self-instruction.

Two self-instruction pilot studies were conducted. The first, in April, 1984, was done to test the success of cognitive modeling and self-guiding statements with little children. Kindergarteners preparing for the Metropolitan Readiness Test were trained in self-instruction and found, through informal observation, to be responsive







-38-


C~~~~~ ~~~ ...~v .,,. yOn* C..... .la


Read each sentence. Then circle the Picture thai matches th sentence.


Si.tyb..u icr i....t i. 'I ad Ft - a t ii.. ci.. un4ii~ Fimbaft by Th.o.ar* CiI
C CflttqAt. ttWtflJ. a, 01a * 0.


1. What hops?


I. Beth con mix i.




2. Ken and Jim have a bol.




3. Now this looks good




4. Ana put a book on the van.




5. Sara is here now.


-k


2. What eats a fish?





3. What digs?


4. What runs?


5. What likes waler?


1 ~ 1


Read each sentence. Then draw a line to the
picture it matches.




1. Jim said. -Who will bake lhi57
Jim said. "Who wil cut this?"



2~ Sara can see the hen.
The hen can see the pig. .


Ken did not help Mom. Ken did help Mom.



The hen cut the bred. Grandma cut the bread.


$


Figure 3-2. Performance Tasks, Set 2 (Clymer et al., 1984,used with permission.)






iC. St�l S 1Stl tm... t. !to. . ej5 lt. *.5 SlktS Cti Stoepto~l e41 . by tCi Orr ,


a fish eo dolphin Look o the f___She Con ploy willi the


rittstplrh tk r t~v*l t flJ' '" *.t.f1 r t
7


Troce the llers to mok words Then write
ore of fhe words 0 h1oisth och senrtentce






1, Q Jim will look in here


j { Hewll see tIle big -!, *H




"You will tile h big whole'
t.{'- Yoo w,l, ___,,_______,_,

3 (. Wil Iru ig whole cot?


/ t thJ wh ........ etl tr,
4 j Will Iht Nit wwii! cll fitt btll?


(r tr ........... ....... w......... .137;


Troco the letters to moe words Then write
one ot tie word, 10 Iish eoch sentlenco






5, i-What book .tt Mom gel?


--4_. _ will got a shork tro.k.
2. tINs Ut0.t oo ,I tot fieth.


- good book It Isl

Who o bgShark this NJstitrd Hetld


"Look al ltis shork, I

4 Wri.hat cn l.t ii. k eo t?


The dolohln can the fish.











Do 1 like titl s

1. This Is _



oo 2 and 3 ltke th is


C





2. Ken get he book, il i








3 She con _ a whole.


lie ~IIl+ .. ... ....
ro- -, dorS ~qli








WithCt tu ri il .-I Where Is t n ieto? ploy?


tftt broll Is hrr ttre ,T'0 ''_..t '. .'




2. Tre COt .... .. not Lire the wate.



3 Wttt ttse Cot rt rn Itt0 --is, 4, Where doesl Ana thke to ploy?



. . . . . . does lth like to ploy?




5. Ano o d ll .. . . the woter.
Alrto u,.d 0ii tt t wl It It11 boll for the Co.


Figure 3-3. Performance Tasks, Set 3 (Clymer et al., 1984, used with permission.)







-40-


/ pi ntt..a far L... ) .iiLl t tf t.Lk


Look at the pictures in each row. Write I in
t picture that comes first.-Write 2 in the
picture. that comes next. Write 3 In the picture
thaot comes lost.







2. Ot Ili 0


0o


-- r m








Read Me Sori Then wite next to mhe first thing thot happens in each stoy. Wnte 2 nt
to the net thring that hoppMn.


1. Gram a will mlt the clov.
Then Grandato will bak" it.


Then Grandmo will b" It.



Grandmo will mix the cloy.

2. Sam and Ken look in.
Sar ond Ken ee Mon with the shark.


Sor on Ken look In.



Saro and Ken see Mom with the shork.


r...Y a t..) -1r i 3~. 171, ..nn.
at)5i~ RA~I~ ~nAM t ~t~~t Ci.ttahrt C~i~t P~tilS. r en cn,./


0 u









0 0





o 00
�6 � �















Look of the pktures in eoch row. Wrfe 1 in the picture that Comes first. Write 2 I oe picture that comes next. Write 3 in the pftice that comes l st.


2.


,. 0 . o


Figure 3-4. Performance Tasks, Set 4 (Clymer et al., 1984, used with permission.)


Ok







-41-


Ch mttt.t ta14/I4. by 0444 *4 iny. 14.

Write a word to finish each sen tece.


1. get gets


Sos .-a book.

2. likes like


The doln

3. works work


4. run runs


An* -- to Jim.

5. se see


Ken he cat.


04.68.0 4.*.~~1


r~Th


r- ,All...aafaaa-t 3.a ft.,aaaaft, aId Read the sentence. Fig In the Ce e t so ft word that goes Into the sentence,


1. Grandma con - bread.
* make 0 nune 0 a


2. A fish - water.
o gets 0 likes 0 need

3. The cot here.
0 lok O o O put

4. Aner and Beth = In the vom
o get 0hops Otooks


5. Betn - for Jim. R
O coMe 0 looks 0 pOls

B Sara - with the cloy.
o ploy 0 worm 0 go


1/A I
Beth alovs with the boll. Beth and Jim ploay with the boll.
work works He with fish. look looks


2. Sara - of dolphin.


Z=:- lke lkes


Jim and Ano the water.


se e..... sees


4. Grandma ________the cat.

U 'ft-


1 1- sei t i~a. by,44 4 C .48 hops digs rum


1. Ken ' A T *" D the von.


eats gets looks


2. The man - a book lor KOO.

looks pks wooks

3. Ken _a the bool.


needs bakes s" es(


4. He a shark In te book


makes puts lies


5. Ken~ ..... the baak


Figure 3-5. Performance Tasks, Set 5 (Clymer et al., 1984, used with permission.)


li !Oil







-42-


fl Ibibiba, la baaai ifllfllfl.4 . bE


Trace the letters to. make words. Then write
one at the words to finish eoch'sentence.






1.-: "-- . What doe the cat need?




2. Ana waft. with clay.


She will make a bell. . He wil bake it now.


He will the fish.

We need a ball.


We a ball to play.


need clay obake 1. You can make a b. bell


2. ereis hatyOU.................



2. Here Is what you 3. Mix the ____4. Work with the -------------5. Now you can the cloy.


ta. 419mw Sloane 'W in~s b mT~aoao1 ci, .. t~



bre has dough


Do I and 3 like this Then do 2 like this -.


e OI744r



2. You bake this. I' 11
Then itlls bread.







3. Jim -bread.W


Figure 3-6. Performance Tasks, Set 6 (Clymer et al., 1984, used with permission.)





-43-


to its structure and format. In May, 1985, a second pilot study was conducted with 10 first grade children to test the organization and choice of materials for this study; the materials were found to be practical and appropriate for use with self-instruction.



Procedure


Subject recruitment. During the preplanning session for the 1985-86 school year, curriculum resource teachers from every elementary school were addressed at their first county-wide meeting and asked to forward to the researcher a report on the number of first graders at their schools and the reading levels to which those children were assigned. Additionally, names of first grade teachers were obtained. Immediately following this report, first grade teachers at each school were contacted by the researcher and invited to participate in the study. This invitation (see Appendix A) included information to the teacher that participation would be relevant to the current emphasis on problem-solving skills development in Alachua County. Nine teachers with a total of 56 students in the beginning reading level of the Ginn series volunteered to participate in the study. Each student/teacher group was randomly assigned to treatment or control groups.





-44-


Subject information collected for analysis included the race, sex, socioeconomic status, and age of each student. Students in the study were not in any special education or remedial education programs or participating in any other educational/tutorial or research programs. The subject recruitment was successful in creating comparable groups of normal first graders ready to be taught to read and perform independently on written tasks.

Teacher training. Training sessions of approximately two hours were scheduled at the convenience of the participating teachers; experimental and control teachers were trained separately. The training for the experiment teachers included (a) information about self-instruction training or the metacognitive approach (see Appendix B);

(b) guidelines for effective self-instruction (see Appendix C); (c) training on student self-instruction training tasks (described below); and (d) all materials for data collection (the training tasks and/or performance tasks). Complete teacher instructions for the experimental group are in Appendix E; for the control group instructions are shown in Appendix F.

Teachers were also provided with a timetable for action which correlated with the Ginn reading pacing.

The training tasks. Self-instruction training tasks should be similar to the performance tasks in terms of their need for strategy selection and implementation, but




-45

should not overtly resemble the performance task. For this reason, non-reading tasks requiring strategy selection and reasoning were selected to train the children in the self-questioning, self-instruction process. These tasks involved matching similar pictures, picture association tasks, completing pathways, tracing figures to completion, reasoning tasks with matching, figure completion, and recognition of quantity. The self-instruction training was done in four separate sessions, and the tasks are illustrated in Figures 3-7 through 3-11.

The experimental groups teachers were trained to use overt-to-covert cognitive modeling and children's speech to formulate four self-guiding questions which included the following steps derived from Meichenbaum's protocol as described in the Introduction and presented in Appendix D:

I. "What should I do here?"

2. "Am I doing what I said I would do?"

3. "What if I make a mistake, how will I fix it?" and

4. "How did I do?" or "Let me check my paper before
I give it to the teacher."

The children developed and practiced the self-guiding

questions on the training tasks, which were broken into four sets of approximately 15-minute sessions. Thus, the experimental group children had about one hour, over four different days, of practice in self-instruction alone.





-46-


Cl 0


Figure 3-7. Training Set 1





-47-


a ..n L


?_ In


g


Training Set 2


- I ft. 7


00
/ ',



/ i


Er


U. a fta


Figure 3-8.





-48-


-. A fl~ a
t~)
S*! A * I

H -___________4 ~ 2. M-'I



_____ I 2.




32.



S.t A ft. S




-'


Figure 3-9. Training Set 3





-49-


Ut 4 1 t


2


Figure 3-10. Training Set 4


en C ~ C


3


z


I ,Q-7


N M





-50-


The children were taught to apply the self-guiding sequence as they worked and to repeat the sequence over again if mistakes were found. These self-guiding questions are, by design and purpose, content-free and later aided the children in selecting strategies from their reading skills repertoires to complete the performance tasks. Teachers guided the children to develop four questions which reflected the mandated steps in the self-instruction model, but understood the importance of accepting the children's wording of those questions. Each experimental group, with students working together, arrived at four questions which were acceptable to the children and met the format.

The performance tasks. The performance tasks selected were the students' independent worksheets from the Ginn skillpack and studybook, as specified in the Instruments section. These follow prescribed lessons which the teachers were encouraged to follow as exactly as possible for the duration of the study. Specifically, for each student, each teacher in the experimental group was provided with

(a) training tasks for self-instruction and (b) a packet of chosen reading tasks keyed to the lesson and point in the teaching series when those tasks should be distributed to the students. The control group teachers received the same materials exclusive of the training tasks.





-51-


Implementation. Ginn Level 3 has a five-week instructional plan. Teachers in both the experimental and control groups instructed reading as usual according to the Ginn teachers' manual. Each group received the same page assignments at approximately the same time. During the third week of instruction in reading, but separately from the reading group setting, the treatment teachers provided four sessions (about 15 minutes each) of selfinstruction training using the provided training tasks. During Level 3 instructional weeks four and five, five minutes of self-instruction review took place in the reading group, and then children were given the performance tasks and data collection began. Again, the self-instruction question set was structured from the previously agreed-upon wording or private speech of each group of children and was content-free; no reference was made to the work at hand, but strategies for doing work were reinforced. The control group participated in reading instruction and received the performance tasks as the experiment group did. All assignments were collected and scored by a master key by the researcher.

All teachers, experimental and control, were in weekly contact with the investigator during the experimental period, and each received a visit during the study. The investigator was easily accessible by telephone or visitation to answer questions and/or model the self-instruction process.





-52-


Statistical Design


Student performance on the Ginnreading tasks, testing the hypothesis that there were no significant differences between experimental and control groups on the reading performance tasks, was analyzed using a multivariate analysis of variance (MANOVA) on a two-way nested design. As shown in Figure 3-11, teachers are nested within group and students are nested within teacher. An alternative design was also considered in which school nested within group was a factor and teachers were nested within school. As subsequent analysis revealed virtually no difference in results for these two designs, the school factor was deleted from the model in the interest of parsimony and ease of reporting results; Figure 3-11 depicts the simpler design for which results are reported in Chapter IV.


GROUP



Teacher Students


EXPERIMENTAL



T T T T S S S S










Figure 3-11. Nesting Design


CONTROL



T T T T T S S S S S





-53


A repeated measures design was considered, but since interest was in comparing overall group performance across tasks rather than group performance by (between) individual tasks, a MANOVA was used.














CHAPTER IV
RESULTS AND CONCLUSIONS


The purpose of this study was to compare the performance of first graders with and without self-instruction strategy training, on academic tasks. Children in the experimental (self-instruction) and control (no selfinstruction) groups each completed six sets of basal seatwork tasks distributed by their teachers to be done independently. Each task set contained a particular set of items from among the basal skills presented for study.

A multivariate analysis of variance (MANOVA) on a

two-way nested design was used to analyze the difference between the two groups for the six tasks. Teachers were nested within group, and students were nested within teachers. Following the MANOVA, six univariate analyses of variance (ANOVAs) were used to compare the difference between the experimental and control group on each separate task. The results of these analyses are presented below.

Table 4-1 presents the means and standard deviations of the different groups on each task.


-54-





-55-


Table 4-1

Means and Standard Deviations for Experimental (N = 29)
and Control (N = 27) Groups


Performance Task Group

Number #Items KR21 Experimental Control

1 21 .88 M 20.55 18.88 SD 0.90 3.65

2 15 .61 M 14.79 14.00 SD 0.67 1.41

3 14 .77 M 13.44 11.25 SD 1.18 2.61

4 10 .36 M 9.62 8.00 SD 0.62 1.17

5 16 .58 M 15.10 12.33 SD 1.01 1.92

6 9 .69 M 8.96 7.85 SD 0.18 1.51




Table 4-2 shows the partial correlation coefficients between tasks 1-6 after controlling for experimental condition. Some tasks have strong correlation and others have little or no correlation; for example, the highest partial correlation observed occurred for tasks 1 and 2; the lowest occurred for tasks 2 and 4.

The MANOVA was performed using SAS and Wilks'

Criterion was used to analyze the overall difference between groups. This multivariate test showed a substantial difference between groups with F = 191.11 with







Table 4-2


Partial Correlation Coefficients,


Tasks 1-6


DF = 53 Task 1 Task 2 Task 3 Task 4 Task 5 Task 6


Task 1
p

Ta sk 2
p

Task 3
p

Task 4
P

Task 5
P

Task 6
p


1.00 0.00

0.44 0.00

0.29 0.02

0.28
0.03

0.08
0.55

0.03
0.80


0.44 0.00

1.00
0.00

0.36
0.00

0.00 0.99

0.37
0.00

0.06
0.66


0.29 0.02

0.36 0.00

1.00 0.00

0.02 0.87

0.22 0.10

0.23 0.08


0.28 0.03

0.00 0.99

0.02 0.87

1.00 0.00

-0.09
0.50

0.27 0.04


0.08
0.55

0.37 0.00

0.22 0.10

-0.09
0.50

1.00
0.00

0.08 0.53


0.03 0.80

0.06 0.66

0.23 0.08

0.27
0.04

0.08 0.53

1.00
0.00





-57-


6 numerator degrees of freedom and 2 denominator degrees of freedom, and p = .0052. Using a pre-established alpha level of .05, this led to a general rejection of the multivariate null hypothesis (at alpha _ .05): There is a difference between the experimental and control groups in terms of performance on the six separate seatwork tasks selected from the basal reader workbooks.

As a follow-up procedure univariate ANOVAs were used to determine if the overall difference between groups existed for each task. Table 4-3 shows the Fs and probabilities for each task.



Table 4-3

Results of Univariate Tests by Task


df Type III SS df Type III SS
(numerator) (denominator) F P Task 1 1 42.15 7 46.59 6.33* .04 Task 2 1 7.56 7 7.29 7.26* .03 Task 3 1 56.74 7 32.94 12.06* .01 Task 4 1 36.72 7 7.13 36.01* .0005 Task 5 1 106.11 7 14.73 50.42* .0002 Task 6 1 11.90 7 19.31 4.31 .07


*Significant for alpha < .05.





-58-


The six individual F-tests using TEACHER(GROUP) mean square error in the denominator (the nesting effect for teacher within group) showed significant treatment effect on each of tasks 1-5, with respective p-values of .04, .03, .01, .0005, and .0002. Although the nesting model was used, the actual effect of nesting was nonsignificant; therefore, these results are conservative.

For task 6, significant nesting effect was found

(p = .008). The F-value for GROUP using TEACHER(GROUP) in the denominator is only marginally lower than the critical F-value required for statistical significance, with p = .07.

To summarize, the effects on five of the six univariate tests are significant. While it is ordinarily good statistical practice to require a higher degree of significance in repeated testing to protect against compounded Type I errors, two aspects of the present study give added protection against compounded Type I error. First, reported univariate F-values used TEACHER(GROUP) as the denominator, even though the nesting effect was not significant. F-tests using residual error as the denominator produced p-values of .017, .016, .0004, .0001, and .0001. The reported F-values are therefore taking a conservative approach and retaining the integrity of the nested model. Second, the MANOVA F was very highly significant (p = .0052), which gives global protection against compounded Type I error of repeated univariate testing.





-59-


To summarize, first graders trained in self-instruction performed significantly better on basal seatwork tasks presented to them to be done independently.

The first task called for word identification in context and reality/fantasy discrimination. This task required matching sentences to pictures and involved reasoning skills. Task 1 was done better by the selfinstruction group.

The second task also called for word identification in context, but with a different format: Children had to circle pictures at the end of each sentence, and there were "distractors" or incorrect pictures available for choice. Children also had to deal with the inflection-s verb skill as they selected their answers. Task 2 was done better by the self-instruction group.

The third task involved recognizing vocabulary words and inserting them into sentences and crossword puzzles. The self-instruction group did better on this task.

On Task 4, reasoning skills were again required. Children had to use numbers to arrange pictures in a logical sequence, and, also, had to read a passage and then sequence its events. The self-instruction group did better than the no-self-instruction group.

Task 5 involved decoding verbs with and without the inflected-s ending and putting those words into sentences





-60-


that made sense. The self-instruction group performed better than the no-self-instruction group on this task.

Task 6 required that students insert vocabulary words into sentences and that they do a crossword puzzle. On this task, on which the self-instruction group performed marginally better (but not to a statistically significant degree), there were two factors which may have confounded the children. One, on the vocabulary-insertion task, some answer words could be used more than once, which broke an implied rule--"use each word once." Two, the crossword puzzle required answers in both the "across" and "down" sets; the children had previously had to insert words in one direction or the other. Each of these differences was subtle. This was the only task on which the nesting effect was significant. It is possible that, perceiving the format difficulties of this task, some teachers extended directions help beyond the requirements of the basal.

One question that occurred as this study was concluded was whether the self-instruction group outperformed the no-self-instruction group by virtue of being better readers: Had something happened which simply increased their reading skill? For this reason, the curriculum resource teachers at each school were again contacted and asked for each child's mastery score on the Ginn reading test for Level 3. This mastery test is teacher-dictated,





-61-


so that the children know, item by item, what to do. There is no problem identification or strategy selection required. On the mastery test there is no significant difference (at a < .05) between the self-instruction and no-selfinstruction groups, as evidenced by the t-test in Table 4-4.



Table 4-4

Scores on Ginn Level 3 Mastery Test



M SD t p



Experimental (N = 29) 44.68 1.94 -1.6 .10 Control (N = 27) 43.62 2.73


Alpha set < .05.















CHAPTER V
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS



Summary


The purpose of this study was to apply the cognitivebehavior modification self-instruction model to classroom practice and content. Cognitive modeling and children's private speech were combined to develop children's selfguiding statements and images as they engaged in independent work from their reading curriculum.

From a review of the literature on the use of

cognitive-behavioral strategies with children, the following research needs were indicated:

1. The use of cognitive-behavior behavior with normal subjects on academic tasks in natural settings must be tested.

2. Some research must focus on the processes rather than the product(s) of learning.

3. The capacity to obtain increased generalization or transfer of learning through cognitive-behavior modification must be tested.

4. There must be greater specificity of the independent variable in cognitive-behavior research.


-62-





-63-


5. The impact of cognitive-behavior modification as a treatment in classroom and home settings has not been tested.

6. There is a need to show that metacognitive changes are reflected in normal performance.

Fifty-six first grade students (within nine intact

reading groups in five schools) were randomly assigned to two treatment conditions. Each reading group was within a normal first grade classroom in a public school; although some schools had more than one experimental group, no school contained both an experimental and control group, and no classroom had more than one participating reading group. There were three urban and two rural schools, with one of the rural schools in each treatment condition.

The treatment condition consisted of self-instruction training using non-reading materials (picture association tasks, mazes, counting tasks) imposed separately from reading instruction. The control group received reading instruction only. The dependent measure for this study was children's performance on the usual independent seatwork tasks provided with the reading series; those tasks were selected and removed from the children's daily workbooks and totaled 85 items in the areas of phonics, word identification in context, and sequencing. The tasks were grouped into six sets for data collection according to the pacing requirements for reading instruction.





-64-


A multivariate analysis of variance (MANOVA) on a two-way nested design was used to compare overall group performance across tasks; this analysis showed a very high difference between experimental and control groups (F= 191.99, p = .0052). Six separate univariate ANOVAs were used to determine if the overall difference between groups existed for each task, and these analyses did show a measurably significant difference between groups on five of the six tasks.


Conclusions

Overall, the results of this study indicate that the

cognitive-behavior modification technique of self-instruction is effective and appropriate as a means of improving the academic task performance of normal children. It is important to note that the positive results here were obtained with groups of children in naturalistic settings and on specific academic tasks done without teacher assistance. For the self-instruction group to do significantly better indicates that cognitive strategy changes can be seen on measures of ordinary performance, and children's use of study time could be improved through development of skills in self-; monitoring.




-65-


Children trained in self-instruction applied academic skills with significantly higher accuracy through the development of planful behavior. Since self-instruction training was implemented with pictorial or reasoning ta-sks rather than reading items, its successful use in the reading area by students reflects its focus on processes or executive skills and demonstrates its generalizability across some tasks.

The results of this study reinforce the appropriateness of the practical application of self-instruction: Teachers were easily and effectively trained to incorporate it into their curricula, and students' performance on academic tasks directly from the district-based curriculum improved.


Implications

The results of this study have implications for

further research on the "pedagogical potential" of cognitive-behavior modification. If self-instruction is instrumental in producing a positive change in the independent work and skills application of young students over a period of a few weeks, its continued (year-to-year) use could possibly establish a repertoire of metacognitive or executive skills students put to use as they work.





-66-


As students improve processes or strategies with

which to approach problems, attendant positive changes in products (test scores, problem solving) should be apparent and might be compared to similar successes attained through methods from the area of metacognition. Also, as executive processes improve, their generalizability and durability over time could be examined.

Eventually, attention could be given to finding the critical elements within the self-instruction process to find those which directly apply to school curricula and/or to certain age groups.

The effective use of self-instruction in a study such as this one but with children over seven remains to be done; once children have begun to internalize speech, adjustments in the type of self-instruction training might have to be made.

Most importantly, the successful results of this

study were due to (a) careful adherence to Meichenbaum's guidelines for effective self-instruction; (b) thorough task analysis and expert knowledge of the students engaged in task performance; and (c) thoughtful, patient attention to the development of cognitive processes. Replication or further application of this work should- cautiously

include those cares.





-67-


For education and for children, successful applied use of self-instruction has other implications. One direct result of self-instruction seems to be students' increased independent behavior by virtue of being able to identify problems better and then plan behavior (or perhaps it is their new knowledge that they are expected to do those things rather than wait for the teacher). Although concern here was for an increase in normal class-task performance, it is logical that such behavior would carry over to tests, homework, and projects assigned to be done independently. It would be ideal for educators to have methods for emphasizing cognitive processes which would, in turn, positively affect the learning products. Educators have spent time and money to raise standards of education through program evaluation, teacher evaluation, examination of pupil/teacher ratios and funding levels; the results that can be obtained with self-instruction suggest that working with students on cognitive strategies may help them make better use of resources currently available.



Recommendations


Given the success of self-instruction on specific

classwork tasks in the reading area by first graders, the obvious expansions of this investigation would be as follows:





-68-


1. Replicate this study with tasks from other content areas.

2. Compare classes with and without self-instruction in a year-long study.

3. Follow classes with and without self-instruction into the next grade, looking for generalization and stability of strategy learning.

4. Train children in self-instruction and investigate their use of it across subject matter and situations.

5. Expand the five-to-seven year old age category

utilizing self-instruction in academics to see if internalization of speech hampers the use of self-instruction or visa versa.

This researcher has particular interest in integrating a full-year, full-class self-instruction regime into the curriculum for a kindergarten group, working long-term to fully automatize the self-instruction strategies, followed

by comparison of those students' work in first grade to students without self-instruction training.














APPENDIX A
INVITATION TO PARTICIPATE



Dear First Grade Teacher,

I am a kindergarten teacher at Stephen Foster and a doctoral student in educational psychology at the University of Florida.

I would like to invite you to be part of the research project for my dissertation; the topic is teaching cognitive strategies to young children. You will find this directly relevant to current interest in problemsolving, developing independent work habits, and teaching thinking skills.

I am looking for teacher volunteers and am primarily interested in using students in Ginn Level 3. Here is a brief overview of what involvement would mean (using Level 3 as an example):

Week 1: Begin Level 3 instruction as usual.

Week 2: Continue Level 3 instruction as usual.

Week 3: All continue Level 3 instruction. Those in
the experimental group will provide four fifteen-minute cognitive strategy lessons
for their students. Those in control groups
will teach as usual.

Week 4: Continue Level 3 instruction as usual.
Experimental group: Distribute the independent seatwork tasks to your students, reminding them that the strategies they
have learned might be good to use. Control group: Distribute the independent seatwork
tasks to your students.

Week 5: Same as previous week.


-69-





-70-


All work for both groups will be provided. That work
comes directly from the Ginn workbooks and will be organized, stapled, and delivered, along with a pacing schedule keyed to the Ginn manual.

You would not need to check any experimental or control work. Errors are expected from both groups. No individual or class data will be used or reported; interest is in comparing the performance of the experimental and control groups which will be comprised of students from several schools.

TRAINING

Before the third week of instruction, teachers
randomly assigned to the experimental group will be trained in teaching cognitive strategies and provided with all materials. Those in the control group will meet (separately from experimental) and receive procedural training and prepared materials.

Training will be fun (!), very informative, and brief (maxirdum two hours for experimental, a little less for control). At the end of data collection, training will be provided for control group teachers and results will be shared with everyone.

The strategies you learn to teach your students will work with all subject areas, are pleasing to the children, and enhance the independent work level in your class.

Please consider joining this study!

Thanks,








If you have Level 3 students in your first grade, and are willing to participate in this study, please notify your curriculum resource teacher or call me at Stephen Foster (372-4363) or at home (378-3992).














APPENDIX B
SELF-INSTRUCTION TRAINING



In self-instruction training, cognitive modeling and private speech are combined as an approach to teaching thinking skills. Children are taught how to use selfstatements and images to think and plan behavior; they are not instructed in what to think or encouraged to focus on "right answers." Although right answers would appear to be a logically correct educational goal, both metacognitive and self-instruction researchers reject that aim and set the development of underlying thinking skills or executive processes as a priority of teaching efforts.

Self-instruction in task performance requires that the student engage in problem definition, appropriate strategy selection, self-evaluation of performance, selfreinforcement, and coping behaviors. To date, the selfinstruction process has been used with young children on social, interactive, motor, and academic tasks. Meichenbaum's model includes the following steps:

1. cognitive modeling (an adult performs the
task and talks to self);

2. child performs the task with model's direction for guidance (overt external guidance);


-71-





-72-


3. child performs task and talks aloud to self
(overt self-guidance);

4. child whispers and performs task (faded,
overt self-guidance);

5. child performs task via inaudible or private
speech or nonverbal self-instruction (covert
self-instruction). (Meichenbaum, 1977,
p. 32)

At step three the child's skills in problem definition, planful behavior, and implementation of strategies can be observed.

Meichenbaum begs the question of self-instruction research in academics by stating,

We can conceive of academic tasks where the teacher provided the children with a set of
tasks . . . and the child's job was to identify
what the problem is, how he or she will go
about solving the task, where the likely pitfalls are, etc. . . . Teachers could give
assignments and ask the children to describe in detail how they are going to go about performing the assignment. . . . Discussion could
center on the process, not only the product,
of the assignment. (Meichenbaum, 1985,
p. 421)

The purpose of this study is to apply the cognitivebehavior modification self-instruction model as developed by Meichenbaum to actual classroom practice and content. Teachers participating in the experimental group of this study will supplement regular reading instruction with students' training in self-instruction. The student training will involve four sessions of approximately 15 minutes during which the teacher will, through cognitive modeling and the use of the children's own speech,





-73-


develop self-guiding statements based on Meichenbaum's model. Using that overt-to-covert style, the self-guiding statements for each group will follow the pattern of

1. problem definition and strategy selection
("Hmmm. What do I do here?");

2. strategy implementation
("Am I doing what I said I would do?");

3. error management
("What if I make a mistake, how can I
fix it?"); and

4. self-reinforcement/self-checking
("How did I do?" or "Let me check my paper
before I give it to the teacher.").

Measures of normal performance will be taken on a set of reading tasks from the Ginn studybook and skillpack. Our interest is in the children's use of selfinstruction, not in individual scores or.teacher performance. Errors are, as usual, expected, and the teacher should feel comfortable in accepting the children's independently done work without correcting it.















APPENDIX C
GUIDELINES FOR EFFECTIVE SELF-INSTRUCTION



Ideally, the teacher should model a metacognitive

perspective, one which indicates openness to thinking and anticipation of mistakes which, when they occur, are viewed as problems to be solved. The instructor should be careful of attributions made when failures occur, with statements directed at nurturing the problem-solving attitude.

To maximize generalization, training in self-instruction must be on a task which has the same requisites or elements of the performance task, but is overtly different from that performance task. The pupil must be a collaborator in generating the strategies to be used, and the private speech trained must be developed with and compatible with the child's natural style. The cognitive strategy training should take place after component skills and experience with the task are in the child's repertoire, and the pupil must be helped to recognize the new task as one facilitated by self-instruction and requiring the transfer of those strategies. The instructor needs to ensure against rote repetition of patterned strategies, encouraging personal involvement by the student by use of


-74-





-75


faded cues and acceptance of the child's wording. The teacher should directly encourage the child to generalize strategies to various tasks or situations, perhaps by engaging in discussion of other suitable uses. Lastly, a sense of self-satisfaction should be nurtured in the child, with attention given to developing the child's self-reinforcing and coping skills as flexibility in the use of strategies is fostered.














APPENDIX D
MEICHENBAUM'S SELF-INSTRUCTION MODEL



Meichenbaum's self-instruction model includes the

following steps:

1. cognitive modeling (an adult performs the
task and talks to self);

2. child performs the task with model's direction for guidance (overt external guidance);

3. child performs task and talks aloud to self
(overt self-guidance);

4. child whispers and performs task (faded,
overt, self-guidance);

5. child performs task via inaudible or private
speech or nonverbal self-instruction (covert
self-instruction). (Meichenbaum, 1977,
p. 32)


-76-















APPENDIX E
INSTRUCTIONS FOR THE EXPERIMENTAL GROUP



Experimental Group

Look over the chart for Level 3 indicating the pages that will be used in the study (attached). Please discard those pages from your student books.

During the third week of instruction, plan four 15-minute Training Sessions in self-instruction. These should be separate from reading group. Do not make any reference to reading.

Training Sessions

The training sessions are used to elicit the four
self-instructional questions from your particular group. Use your reminder card (pictured below) for guidance in structuring the questions, but accept your own children's wordings. Work to get four specific questions, in general langauge (not a rote phrase) that appeals to your group.

Introduce this in the following way:

"When you are in school you have a lot of jobs
to do. Sometimes you need help with directions.
Other times you can figure out what to do by
yourself. Let's start a Thinking Class together,
and look at some work to see if, together, we
can ask some questions that will help you to
figure out what to do on your own."

Use Training Set 1

What is the problem on the page?

How can you mark the answer?

Sometimes when you are working, the teacher comes
around and checks to see if you are working correctly. Is that something you could do for
yourself?


-77-





-78-


How?

What if you make a mistake?

What would people who are thinking want to do?

What usually happens when you turn your work
in to the teacher?

What does she find?

Is that something you could do for yourself? Teacher guidance and summation:

Let's go back and see what happens in Thinking
Class. First you found the problem on the page-is there a question you can think of to help
find the problem?

"Hmmm. What do I do here?"*

After you figure out what to do, how can you be your own teacher and see how your work is going?

"Am I doing it?"*

What if you make a mistake, what can you do to
fix it and make it OK with yourself?

"Am I messing up? What can I do?"*

What is a question you could ask yourself in order to be your own "checker" so the teacher won't find
so many mistakes?

"How did I do?" "Let me check this."*




*Samples of answers developed by other students in selfinstruction. Your students' self-guiding statements or questions will reflect their mutual language and style. Repeat the process, streamlining it, over the next three training sessions.





-79-


Data Collection

After the last training session, in a reading group time, give children Set 1 of the performance tasks. To receive this set, they should be through the story "Look at Fish." Spend a minute remininding them of self-instruction questions before the set is distributed. Have them do Set 1 independently; take up the work and go on with reading group as you wish. Do not help the children with their independent work or worry about their mistakes.

Continue instruction according to the Ginn manual. There are five data collection points left. Fit them in in the following order:

After "What Grandma Does"--Set 3

After "Mix and Make"--Set 5

After "Bread to Eat"--Sets 4 and 6 (in two separate
sessions)

After "The Hen and the Bread"--Set 2

Put each child's work in a pile of six sets and call me up! I will take it away and see what happened and will be back in touch with you.


Thanks,







Additional Administrative Procedures

I will place a call to you weekly to check your pacing and progress. I will ask for your training session times and your reading group times for the upcoming week, and answer any questions you have. Additionally, observation times during the term of the study will be planned, and I may drop in at the training session or reading group times.





-80-


Reminder Card

(You will have this card, laminated, to keep with you to channel your students' questions.)


PROBLEM DEFINITION AND STRATEGY SELECTION
"Hmmm. What do I do here?"

IMPLEMENTATION
"Am I doing what I said I would do?"
"Am I doing it?"


ERROR MANAGEMENT
"What if I make a mistake, how can I
"Am I messing up?"
"Ooops! I can cry or I can fix it!"

SELF-REINFORCEMENT/SELF-CHECKING
"How did I do?"
"Let me check my paper before I give


fix it?"


it to her!"





-81-


Level 3 Pacing Sheet

Here is a list of the stories from Level 3. The skillpack
(SP) and studybook (SB) pages I am interested in are underlined.


Text

6-9


Story


Ken and the Fish


10-13 Ana and the Whale 14-19 What Is It? 20-21 Here Come the Whales


SP 1 SB3
SP 7 SB 9
SP 13 SB 16


22-26 Look at Fish


SP 19 20 SB 23 24


21 22 23 24 25 25 26 27 28 29


30-35 What Grandma Does 36-43 Mix and Make 44-48 Clay to Bake 49-55 Bread to Eat 56-63 The Hen and the
Bread


SP 25 26 SB 30 31
SP 32 33 SB 37 38 SP 39 40
SB 43 44 SP 46 47 SB 50 51 SP -5 3 54 SB 57 58














APPENDIX F
INSTRUCTIONS FOR THE CONTROL GROUP



Control Group

Look over the chart for Level 3 indicating the pages that will be used for the study (attached). Please discard those pages from your student books.

Data Collection

After finishing the story "Look at Fish" distribute
task Set 1 to your students to be done independently. When they finish, take up the work and go on with reading group as you wish. Do not help the children with their independent work or worry about their mistakes.

Continue instruction according to the Ginn manual. There are five data collection points left. Fit them in in the following order:

After "What Grandma Does"--Set 3

After "Mix and Make"--Set 5

After "Bread to Eat"--Sets 4 and 6 (in two separate
sessions)

After "The Hen and the Bread"--Set 2

Put each child's work in a pile of six sets and call me up! I will take it away and see what happened and will be back in touch with you.

Thanks,


Additional Administrative Procedures

I will place a call to you weekly to check your pacing and progress. I will ask for your reading group times for the upcoming week and answer any question you have. Additionally, observation times during the term of the study will be planned; I may drop in at the reading group times.


-82-





-83-


Level 3 Pacing Sheet


Here is a list of the stories from Level 3. The skillpack
(SP) and studybook (SB) pages I am interested in are underlined.


Text


Story


6-9 Ken and the Fish SP 1 SB3
10-13 Ana and the Whales SP 7 SB 9
14-19 What Is It? SP 13 SB 16
20-21 Here Come the Whales


22-26 Look at Fish


SP 19 20 SB 23 24


21 22 23 24 25 25 26 27 28 29


30-35 What Grandma Does 36-43 Mix and Make 44-48 Clay to Bake 49-55 Bread to Eat 56-63 The Hen and the
Bread


SP 25 26
SB 30 31 SP 32 33 SB 38 38 SP 39 40 SB 43 44 SP 46 47
SB 50 51 sP -3 54 SB 57 58


29 30
34 35 36 37
41 42 43 44 47 48 50 51 54 55 57 58 61 62













APPENDIX G
RESEARCH DESIGN


E1 E 2

X 1 X 2 X 3 X 4 X 5 X 6 X X2 X3 X4 X5 X6



s6 s 40 T2 7 T6 141
I I



S16 43 T3 S17 T7 S44

I I

$ 24 $ 48 T 4S 25 T8 S49
I I



S 35 S57



T9 P58
I

$63


Where X1 ...... X6 are the six task scores for each child

S1 ...... S63 are the children (subjects)

T1 ...... T9 are the teachers

EI-E2 are the experimental and control conditions


-84-














REFERENCES


Bandura, A. (1969). Principles of behavior modification.
New York: Holt, Rinehart and Winston.

Brown, A. (1980). Metacognitive development and reading.
In R.J. Spiro, B.C. Bruce, & W.F. Brewer (Eds.),
Theoretical issues in reading comprehension: Perspectives from cognitive psychology, linguistics,
artificial intelligence and education. Hillsdale,
New Jersey: Lawrence Erlbaum Associates.

Clymer, T., Venezky, R.L., Johnson, D.D., & Pearson, P.D.
(1984). Fish and not fish. Lexington, Massachusetts:
Ginn and Company.

Craighead, W.E., Craighead-Wilcoxon, L., & Meyers, A. (1978).
New directions in behavior-modification with children.
In M. Hersen, R.M. Eisler, & P.M. Miller (Eds.),
Progress in behavior-modification, Vol. 6 (pp. 159-201).
New York: Academic Press.

Craighead, W.E. (1982). A brief clinical history of
cognitive-behavior therapy with children. School
Psychology Review, 11(l), 5-13.

Craighead, W.E., Meyers, A.W., Craighead, L.W., & McHale,
S.M. (1983). Issues in cognitive behavior therapy with
children. In M. Rosenbaum, C.M. Franks, & Y. Jaffe (Eds.), Perspectives on cognitive therapy in the eighties (pp. 234261). New York: Springer.

Gagne, R. (1964). Problem solving. In A. Melton (Ed.),
Categories of human learnings (pp. 293-317). New
York: Academic Press.

Grimm, J., Bijou, S., & Parsons, J. (1973). A problemsolving model for teaching remedial arithmetic to
handicapped young children. Journal of Abnormal
Child Psychology, 1, 26-39.


-85-





-86-


Hobbs, S., Moguin, L., Tyroler, M., & Lahey, B. (1980).
Cognitive-behavior therapy with children: Has
clinical utility been demonstrated? Psychological
Bulletin, 87, 147-165.

Kazdin, A. (1982). Current developments and research
issues in cognitive-behavioral interventions: A commentary. School Psychology Review, 11(l), 76-82.

Kendall, P. (1977). On the efficacious use of verbal
self-instruction procedures with children. Cognitive
Therapy and Research, 1, 331-341.

Lovitt, T., & Curtis, K. (1968). Effects of manipulating
antecedent event on mathematics response rate.
Journal of Applied Behavior Analysis, 1, 329-333.

Meichenbaum, D. (1977). Cognitive-behavior modification:
An integrative approach. New York: Plenum Press.

Meichenbaum, D. (1985). Teaching thinking: A cognitivebehavioral perspective. In S. Chipman, J. Segal, &
R. Glaser (Eds.), Thinking and learning skills
(Vol. 2): Research and open questions (pp. 407-426).
Hillsdale, New Jersey: Lawrence Erlbaum Associates.

Meichenbaum, D., & Asarnow, J. (1979). Cognitive-behavior
modification and metacognitive development: Implications for the classroom. In P. Kendall & S. Hollon
(Eds.), Cognitive-behavior interventions: Theory,
research and procedures (pp. 11-35). New York:
Academic Press.

Meichenbaum, D., & Burland, S. (1979). Cognitive-behavior
modification with children. School Psychology Digest,
8(4), 426-433.

Meichenbaum, D., & Burland, S. (1981). Cognitive-behavior
modification with children. In H.F. Clarizio, R.C.
Craig, & W.A. Mehrens (Eds.), Contemporary issues in
educational psycholoqy (4th Edition) (pp. 109-114).
Boston: Allyn and Bacon.

Meichenbaum, D., & Goodman, J. (1971). Training impulsive
children to talk to themselves: A means of developing
self-control. Journal of Abnormal Psychology, 77, 115-126.

Meyers, A.W., & Craighead, W.E. (1984). Cognitive behavior
therapy for children. New York: Plenum Press.





-87-


Pressley, M., & Levin, J.R. (1983). Cognitive strategy
research: Educational applications. New York:
Springer-Verlag.

Rhodes, D. (1979). Cognitive self-instruction to increase
comprehension in early readers (Doctoral dissertation,
University of Louisville, 1979). Dissertation
Abstracts International, 41/02A, 612.

Sullivan, F.J. (1981). A comparison between attack
strategy training and attack strategy training in
combination with self-instruction in teaching
academic tasks to first graders (Doctoral dissertation, University of Oregon, 1981). Dissertation
Abstracts International, 42/08A, 3520.

Vygotsky, L.S. (1962). Thought and language. Cambridge,
Massachusetts: Massachusetts Institute of Technology
Press.














BIOGRAPHICAL SKETCH


Christopher Williams Sparks was born on March 23,

1946, in Shreveport, Louisiana, second in a family of six children. She grew up in university towns and settled in Gainesville, Florida, as a result of her family's migration there.

Chris has a Bachelor of Arts in Education and a Master of Education from the University of Florida, with specialization in early childhood. She will receive a Doctor of Philosophy in foundations of education in May, 1986, from the University of Florida.

Chris has been a teacher in the Alachua County, Gainesville, Florida, school district for 18 years. Although she has a preference for kindergarten-age people, she has also enjoyed her part-time employment in the Childhood Education department at Gainesville's Santa Fe Community College.

Post-graduation plans for Chris include further

research in cognitive-behavior modification, publication of several articles which have been on the "back burner," and continued teaching. Additionally, she intends to reacquaint her patient children Judd and Sara, who have


-88-





-89



been eating a lot of frozen pizza, with good food and motherhood.










I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Gordon E. Greenwood-, Chairman Professor of Foundations of Education







I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Johh M. Newell
Professor of Foundations of Education







I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Linda M. Crocker
Professor of Foundations of Education










I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Suzanne L. Krogh / Associate Professor of Lnstruction and Curriculum







This dissertation was submitted to the Graduate Faculty of the College of Education and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy.

May 1986


Chairman, Foundations of Education




Dean, Colleg of Educa ion


Dean, Graduate School




Full Text

PAGE 1

COMPARISON OF SPECIAL EDUCATION LABELING PRESCHOOL VERSUS ELEMENTARY SCHOOL AGE CHILDREN BY PATRICIA ANN MAHER A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1982

PAGE 2

ACKNOWLEDGEMENTS During the planning, development and completion of this dissertation, the guidance and encouragement of Dr. Arthur Lewis, committee chairman, have been greatly appreciated. His wide range of interests and acceptance of new ideas foster creativity and a spirit of independence which insure that a dissertation becomes a basis for new beginnings as well as a culmination of past efforts. Dr. Robert Soar served on the committee as research advisor and offered invaluable assistance as the methodology and analysis portions of the study were developed. Dr. Barry Guinagh also served as a committee member. His assistance was especially helpful as the topic of this study was developed. He understood the subtle power and far reaching effects of special education labels. Dr. Robert Lange, of the University of Central Florida, gave generously of his time and expertise in the design and interpretation of the computer analysis. Because of his patience, this portion of the study became interesting, challenging and enjoyable. Throughout the extended time involved in the completion of the dissertation, the interest, concern and encouragement of family, friends and colleagues have been a continuous source of strength and renewal. I am especially indebted to Teri Maher who offered special assistance ii

PAGE 3

in preparation of the data analysis cards, to Shelby Morrison who encouraged me to begin this endeavor, to Pat Morgan whose insistence on meticulous records provided much of the data utilized for the study, and to Patty and Worth Dalton who provided the insight of parents of young children and believed the topic of this study was important. iii

PAGE 4

TABLE OF CONTENTS ACKNOWLEDGEMENTS ii ABSTRACT vi CHAPTER PAGE ONE THE PROBLEM 1 Problem Statement 1 Justification for the Study 2 Definition of Terms 7 Design of the Study 8 Organization of the Report 11 TWO RELATED LITERATURE 12 Introduction 12 Theoretical Basis for Categorization and Labeling 12 Historical Review 14 Expectancy Effects 18 Funding Influences 29 THREE PROCEDURES 32 Introduction 32 Design of Research 33 Data Collection 35 Assumptions and Limitations of Research 41 Data Analysis 41 FOUR ANALYSIS AND PRESENTATION OF RESULTS 45 Analysis of First Hypothesis 45 Analysis of Second Hypothesis 49 Analysis of Third Hypothesis 55 FIVE SUMMARY, MAJOR FINDINGS, DISCUSSION, SUGGESTIONS FOR FURTHER RESEARCH 79 Summary 79 Major Findings 80 Discussion 82 Suggestions for Further Research .... 87 Conclusion 90 iv

PAGE 5

APPENDIX A DISCRIMINANT FUNCTION ANALYSIS STEPWISE PROCEDURE 93 APPENDIX B DISCRIMINANT FUNCTION ANALYSIS STEP BACKWARD PROCEDURE . . 95 REFERENCE NOTES 97 * REFERENCES HH BIOGRAPHICAL SKETCH 105 V

PAGE 6

Abstract of Dissertation Presented to the Graduate Council of the Ur.iversity of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARISON OF SPECIAL EDUCATION LABELING IN PRESCHOOL VERSUS ELEMENTARY SCHOOL AGE CHILDREN By Patricia Ann Maher May, 1982 Chairman: Dr. Arthur J. Lewis MEjorDepartment: Curriculum and Instruction Howard Becker's theory of social deviance established the importance of labels, applied to persons, in creating and maintaining deviant status by influencing the expectancies of significant persons within the social system. Because of the effects of special education classifications (labels) on the expectancies of the child, paren's, teachers and others, the application of these labels becomes a decision with enduring consequences. The purpose of this study was to investigate differences in the stability of special education labels applied to a preschool and an elementary school age group. Specifically, the study sought to determine if labels applied to preschool children changed more frequently than labels applied after children had begun elementary school. Additionally, the relative stability among the categories, educable mentally retarded (EMR), trainable mentally retarded (TMR) vi

PAGE 7

and specific learning disabled (SLD), was examined. And, finaily, the variables of grade (preschool or elementary), initial label, sex end race vvere explored in regard to label change. Students served by the Orange County School Board in centra! Florida were the subjects of this study. Data were accumulated on 298 children referred for reevaluation after placement in special edijcavion programs as preschool or elementary school students. Analysis was accomplished through calculation of contingency tab'es, ccrrelal.ion coefficients and application of discriminant function analysis. The findings of this study showed that labels changed significantly more frequently when applied to preschool than to elementary school childrr-rn Among the categories included in the analysis, the EfVIR label shewed the most instability. When interactions .^mong independent varic-bk.'s were examined, the effect of the grade at the tiire oT initial classivication and the type of label applied outwaigheci oxher interactions. When the various subgroups were compared, it was determined that preschool, white, EMR males changed labtis rnce frequently chsn any other subgroup. The results of this study suggested that, because of the instability of l^ibels applied to preschool children, classific=3t!on procedures for this group should be carefully reviewed and alternatives such as noncategorical programs should be investigated. vii

PAGE 8

CHAPTER I THE PROBLEM Problem Statement Diagnostic classification prior to special education placement has frequently been criticized as a potentially harmful process. Researchers have found that diagnostic labels convey powerful meanings that may influence a teacher's understanding of and response to a child (Algozzine, Mercer & Countermine, 1977; Carroll & Reppucci, 1978; Seitz & Geske, 1976). Tuckman (1972) stated that "in addition to lowering teacher expectations, the label creates an emotional upheaval, degraded self-image, and self-fulfilling prophecy of lowered potential" (p. 169). Because of these expectancy effects, the accuracy of initial classifications becomes crucial. If the label is inaccurate or unstable, the expectancy effects may remain even if the label is later corrected. The purpose of this study was to investigate the relative stability of special education labels applied to preschool children and to elementary school age children who had attended special education programs and had been referred for reevaluation at teacher request or as part of the required three year reassessment. The problem under consideration was one that could best be studied using experimental methods. However, that would have necessitated identification of a group of handicapped children and assignment of one section to a special education program and assignment of the other section to a control condition of "no placement." Since this study dealt with human 1

PAGE 9

2 subjects, denial of appropriate educational placement involved serious ethical considerations. Because of these ethical constraints, the experimental method was not applied. The study relied upon data obtained from real children placed in actual special education programs in one county school system. Justification for the Study Subsequent to the enactment of the Education for All Handicapped Children Act of 1975, PL 94-142, the number of handicapped preschool children enrolled in special education programs increased nearly 20% nationally from 196,223 in 1976 to 231,812 in 1979 (WESTAR, 1981). This law mandated that a free, appropriate, public education be provided for all handicapped children between the ages of three and twenty-one (Abeson & Zettel, 1977; Cole & Dunn, 1977; Powell, 1978; "Public Law 94-142," 1981; Shanker, 1980). its provisions brought special attention to the population of three to five-year-old children, a group previously considered the responsibility of parents and private preschool programs and, therefore, unserved by public schools (Cohen, Semmes & Guralnick, 1979). Two sources of funding for preschool programs were provided by PL 94-142, state entitlement monies and preschool incentive grants. State entitlement money was allocated according to the number of children identified, classified and served in special education programs. Preschool children generated the same amount of money as special education children six years of age and older according to the appropriation formula of the law. The second source of funding.

PAGE 10

preschool incentive grants, was established as an amendment to the law to encourage states to develop preschool programs. Although PL 94-142 included the three to five year old group in its provisions, it exempted states whose statutes were not consistent with the law. That is, those states whose educational practices designated five as the entry school age were not bound to the requirements of the law. In recognition of the importance and cost effectiveness of early intervention with handicapped children, preschool incentive grants were established to encourage states to develop educational programs for this age group. The preschool incentive grant program provided funds to initiate, improve and expand services for these young children (Cohen et al . , 1979). The emphasis on programming for handicapped preschoolers offered the advantage of early intervention and the possible prevention of more serious learning problems. However, because funding was accomplished categorically, it was required that children be tested, diagnosed and labeled before being assigned to programs. While this process may have been fiscally effective, its application to a preschool population was limited by the unique characteristics of this group. Dangers in early identification and placement were related to difficulties of assessment, variabilities in child development, and possibilities of misdiagnosis and inaccurate categorization (Mercer, Algozzine & Trifiletti, 1979).

PAGE 11

Assessment of young children is limited by variations in behavior and attention. Distractibility , short attention span and separation anxiety may influence a child's performance and limit the accuracy of the evaluation (Cataldo, 1981). Many of the tests used to identify disabilities are somewhat unreliable with young age groups (Divoky, 1974). Keogh (1977) stated that children may be identified as "at risk" on the basis of inconclusive, fragmentary and selective test results. Identification and diagnosis of problems such as mild retardation have been less accurate than identification of more severe handicapping conditions (Aldrich & Holliday, 1971). Definitions of disability categories are often vague and may differ from state to state. An analysis of state definitions of preschool handicapped programs showed that only seven states of forty-four surveyed had determined definitions specific for the preschool population (Lessen & Rose, 1980). Disability categories have been designed for school age populations and often are not appropriate for preschool groups. Keogh (1977) described early identification as a predictive activity. She stated, "We are trying to identify children who are likely to have later school problems, even though they have not yet been exposed to formal school programs" (p. 268). Another problem of special significance to preschool groups is the variability in normal development among children. The influence of environmental experiences may intensify maturational differences during early years, and large discrepancies in developmental rates can be observed (Cataldo, 1981). Behaviors, speech and language skills and

PAGE 12

5 motor skills can change rapidly and may confuse a diagnostic picture. A temporary developmental delay may be considered a major problem if it is observed only during a single testing session (Mercer et al., 1979). The combination of measurement inadequacies and differential development complicates the assessment of children and increases the possibility of misdiagnosis with resultant incorrect or unstable classification. This is the major disadvantage of early identification. Historically, the development of categories for exceptional children was based on tenets of help and service (Hobbs, 1975b). However, when the categories of exceptional children are examined, it is evident that the labels imply these children are "exceptional" because they create problems which are considered undesirable, harmful, or otherwise detrimental to the individual in question and to the welfare of society in general (Kitsuse & Spector, 1973). These individuals must be isolated or treated in a special manner to remedy their exceptionality. Becker's (1963) sociological deviance theory suggests that labels cause changes in the expectancies of significant persons in the social system. The label and related expectancies become the basis for the deviant status which, in turn, is communicated to the person and then reflected in the behaviors of the labeled person and the behaviors of those significant others who interact with him. That is, labels create self-fulfilling prophecies (Erikson, 1962; Goffman, 1974; Matza, 1969).

PAGE 13

6 Concerns over special education labels and expectancy effects have become important issues in special education over the past decade. In spite of contradictory claims presented by a few researchers (MacMillan, Jones & Aloia, 1974), evidence has mounted that special education classifications are responsible for humiliation, stigma, lowered performance and reduced opportunities for the labeled individual (Dunn, 1968; Mercer, 1977). Several researchers have confirmed that diagnostic classifications bias teachers' attitudes toward children (Algozzine et al., 1977; Foster, Schmidt & Sabatino, 1976). The harmful effects of labels may be magnified when applied to a preschool population. Since categorization represents deviancy from the norm and may produce negative expectancy effects, the stability of the label over time is crucial. If a child is designated as retarded when he is three years old and reclassified as learning disabled or normal when he is five years old, simply changing the label may not remove attitudes and expectancies which have developed at school and within the family constellation at home. If classifications are not stable within the preschool group, then alternative methods of placement should be considered. The purpose of this study was, first, to investigate the stability of special education classifications as applied to preschool versus elementary age children; second, to examine the type of label changes which occurred and to compare the frequency of label change among the categories--educable mentally retarded (EMR), trainable mentally retarded (TMR) and specific learning disabled (SLD); and

PAGE 14

7 finally, to investigate the relationship between change of label and grade group (preschool or elementary), initial classification, reevaluation classification, sex and race of the child. Definitions of Terms PL 94-142 . The Education for All Handicapped Children Act of 1975 which mandated a free, appropriate public education for all handicapped children. Specific Learning Disabled . One who exhibits a disorder in one or more of the basic psychological processes involved in understanding or using spoken or written language. These may be manifested in disorders of listening, thinking, reading, talking, writing, spelling, or arithmetic. They do not include learning problems which are due primarily to visual, hearing, or motor handicaps, to mental retardation, to emotional disorders, or to environmental deprivation (Orange County Public Schools, Note 1). Educable Mentally Retarded . One who is mildly impaired in intellectual and adaptive behavior and whose development reflects a reduced rate of learning. The measured intelligence of an educable mentally retarded student generally falls between two and three standard deviations below the mean, and the assessed adaptive behavior falls below age and cultural expectations (Orange County Public Schools, Note 1 ) . Trainable Mentally Retarded. One who is moderately or severely impaired in intellectual and adaptive behavior and whose development reflects a reduced rate of learning. The measured intelligence of a

PAGE 15

8 trainable mentally retarded student generally falls between three and five standard deviations below the mean, and the assessed adaptive behavior falls below age and cultural expectations (Orange County Public Schools, Note 1). Label . To put in a certain class; to classify. (The Random House Dictionary of the English Language , 1966). Expectancy . Expectations about conscious or unconscious evaluation which one person forms of another which leads the evaluator to treat the person evaluated as though the evaluation were correct (Finn, 1972). Deviance Theory . Social groups create deviance by making the rules whose infraction constitutes deviance and by applying these rules to particular people and labeling them as outsiders (Becker, 1963, p. 9). Design of the Study Assumptions The assumptions in this study were: 1. Children were assigned special education classifications based on standardized assessment procedures. 2. Children placed in special education programs received an annual review according to specified county procedure (Orange County Public Schools, Note 1). Subsequent to this review, they were referred for reevaluation and a change of classification and program placement was recommended, if appropriate.

PAGE 16

3. Class records, psychological services records, and computerized student basic identification information were accurate and complete. 4. Both age groups in the study were enrolled in special education programs which were appropriate for their learning problems. Methodology This investigation utilized a correlational design and, therefore, did not presume to assign causality (Campbell and Stanley, 1963). The study dealt, not with contrived situations, but with actual data as they existed within one county school system. The intent of the study was to explore the following questions: 1. Were labels applied to preschool children more unstable than labels applied to elementary age children? 2. Were some classifications more unstable than others? For the purpose of this study, the classifications under consideration were educable mentally retarded, trainable mentally retarded and specific learning disabled. 3. Were the label changes which occurred within the two groups influenced by time of initial classification, type of label applied, individual characteristics such as the sex or race of the child, and factors such as level of parent education and single or dual parent homes? 4. Were the label changes influenced by interactions between the grade group (preschool or elementary) at the time of initial testing, type of label applied, and the sex or race of the child.

PAGE 17

10 The hypotheses tested were: H 1: Among children who have attended a special ° education class and have been referred for reevaluation, there is no significant difference in frequency of label change between those younger than six and those six years old or older at the time of their initial evaluation. H 2: There is no significant difference in frequency of ° label change among the three categories: educable mentally retarded (EMR), trainable mentally retarded (TMR), and specific learning disabled (SLD). H 3: There is no significant variation in classification change attributable to the variables of sex, race, initial classification and grade group (preschool or elementary) or interactions among these variables. The data were collected according to the following procedures: 1. Records from the Orange County Preschool Diagnostic Center were reviewed on children referred for reevaluation after placement in preschool SLD, TMR or EMR programs. For each child the following data were accumulated: sex, race, entry IQ, entry classification, present IQ, present classification, parent educational level and number of parents in the home. 2. Records of all children referred for reevaluation after placement in self-contained classes for elementary age children at the special school for TMR, self-contained classes for elementary age students at the special school for SLD and students placed in self-contained classes for EMR were reviewed. EMR classes are placed in regular schools and records from seven schools were reviewed. For each child the following data were accumulated: sex, race, entry classification, reevaluation IQ,

PAGE 18

11 reevaluation classification, parent educational level and number of parents in the home. The records for the groups extended over a four year period . Statistical treatment of the accumulated data included determination of contingency tables and calculation of correlation coefficients between label change and each of the independent variables. Multivariate analysis was accomplished through application of discriminant function analysis. Limitations 1. This study was limited by loss of records on those students initially placed in programs who moved out of county before receiving their three-year reassessment. 2. This study conformed to a correlational design and did not assume the ability to assign causality. 3. The study was limited to students served by the Orange County School System, Department of Exceptional Education. Organization of the Report The report was organized as follows: Chapter II included a review of related literature; Chapter III consisted of a description of the methodology of the study; Chapter IV included a presentation and analysis of the results of the study; and Chapter V contained a summary of the study, discussion of the findings, and recommendations for additional research.

PAGE 19

CHAPTER II RELATED LITERATURE I ntroduction Literature was reviewed in four major areas: (1) theoretical basis for categorization and labeling, (2) historical development of classification procedures in the United States, (3) expectancy effects and their relationship to labeling and (4) funding influences on categorization policies. The theoretical basis for categorization and labeling was related to the sociological theory of deviance. The historical review focused on the development of practices and procedures of classification in the United States from the 1800s to the present. Expectancy effects were reviewed in regard to their influence on teacher-student interaction and their particular application in the use of special education labels. Finally, funding influences related to the effect of methods of appropriation of funds on the continued use of special education labels. Theoretical Basis for Categorization and Labeling Categorization of individuals or groups according to characteristics which make them different from the norm is based on sociological deviance theory as proposed by Howard Becker (1963). He stated: Social groups create deviance by making the rules whose infraction constitutes deviance and by applying those rules to particular people and labeling them as outsiders. From this point of view, deviance is not a quality of the act the person commits, but rather a consequence of the application by others of rules and sanctions to an "offender." The deviant is one to whom the label has successfully been applied: deviant behavior is behavior that people so label, (p. 9) 12

PAGE 20

13 Becker's deviance theory emphasizes the important effects of labeling as part of the process of establishing deviance. His theory suggests that labels cause changes in the expectancies of significant persons in the social system. Becker further clarified: "Whether a given act is deviant or not depends, in part, on the nature of the act (that is, whether or not it violates some rule) and, in part, on what other people do about it" (p. 14). Edwin Shur (1971) explained deviance theory by providing a "working definition." He stated: Human behavior is deviant to the extent that it comes to be viewed as involving a personally discreditable departure from a group's normative expectations and it elicits interpersonal or collective reactions that serve to isolate, treat, correct or punish individuals engaged in such behavior, (p. 24) Both of these definitions emphasize that deviance is assigned because the individual is somehow different from the norm. This carries the implication of threat to those within the normal group and leads to the consideration of social control. Deviance and control are related because society must be protected from those individuals labeled as deviant. Hobbs (1975a) stated: Classification serves to maintain the stability of the community and of its institutions, to control the allocation of resources and govern access to them, to reduce discord in school and neighborhood, to preserve majority values and expectations, and to allay the anxiety generated by the presence of the deviant individual, (p. 20)

PAGE 21

14 Erikson (1966) defined deviance and control as part of a continuing concept. He explained deviance as: "conduct which is generally thought to require the attention of social control agencies that is, conduct about which something should be done" (p. 308). The issue of social control as it relates to deviance theory explains the establishment of institutions to house the deviants. Within our public schools we have created special classes to provide for exceptional children and, in this context, have established a means of social control for these groups. The term "exceptional children" implies that the group so designated is somehow different from normal children. Kitsuse and Spector (1973) explained that populations of exceptional children were separated due to the existence of certain social conditions. These were attributes which were considered undesirable, harmful, or otherwise detrimental to the individual in question and the welfare of society in general. Consideration of categories of exceptional children reveal terminology such as learning disabled, mentally retarded and emotionally disturbed. These conditions carry connotations of less than normal functioning and clearly refer to groups of children who represent social problems (Hobbs, 1975b). Historical Review Definitions of deviance which are applicable in the United States today have their base in historical events (Hobbs, 1975a). During the 1800s and 1900s heavy immigration from Ireland and southern and eastern Europe brought settlers into the United States who were poor

PAGE 22

15 and uneducated. These newcomers contrasted sharply with the more established Protestant, Anglo Saxon citizenry. Their presence was considered a potential source of economic drain and a threat to the well being of the community. This immigrant group fit the requirements of deviance and providing a measure of social control became important. One method of insuring that they would assimilate American culture rather than cling to old customs was to insist that all children attend the same school system. The enactment of the compulsory school attendance laws in the mid 1800s served the major purpose of providing control over the socialization of this immigrant group. As a result of compulsory attendance, all children were required to attend school and differences in academic progress among the ethnic groups quickly became apparent. Many children were unable to keep pace with the norm in academic achievement, and special educational provisions were needed for them (Hobbs, 1975a). During the same time period, the theories of natural selection and survival of the fittest were introduced. The eugenics movement, which emphasized the influence of heredity on the future strength of the human race and insisted that society should prevent the propagation of the unfit, became popular as a consequence of these theories. This movement was supported by respected scientists of the day including Henry Goddard, Robert Yerkes and Louis Terman (Kamin, 1975).

PAGE 23

16 In 1917, in an article, "The Menace of Feeble-Mindedness, " Terman issued a stern warning: Only recently have we begun to recognize how serious a menace it is to the social, economic and moral welfare of the state ... it is responsible . . . for the majority of cases of chronic and semi chronic pauperism, organized charities often contribute to the survival of individuals who would otherwise not be able to live and reproduce . . . if we would preserve our state for a class of people worthy to possess it, we must prevent, as far as possible, the propagation of mental degenerates, (p. 161) The religious doctrine of predestination was also a forceful influence of the time (Erikson, 1966). According to this doctrine, worldly success was an external sign of an individual's predestination to be saved. Conversely, those not saved were identified by their poverty, ignorance and lack of ambition. Their condition was the result of their inability or unwillingness to be thrifty and their incapacity to use self-denial. The poor were considered disgraceful and relief measures were seen as encouraging weakness of character (Hobbs, 1975a). Those social influences established boundaries of normalcy and deviance and, therefore, defined limits of behavior for society. Erikson (1966) wrote: The deviant is a person whose activities have moved outside the margins of the group, and when the community calls him to account for that vagrancy, it is making a statement about the nature and placement of its boundaries. It is declaring how much variability and diversity can be tolerated within the group before it begins to lose its distinctive shape, its unique identity-deviant forms of behavior, by marking the outer edges of group life, give the inner structure its special character and thus supply the framework within which the people of the group develop an orderly sense of their own cultural identity, (pp. 11-13)

PAGE 24

17 Categorization identified deviant members of society and separated them from normal individuals. Among these groups were children who did not fit with regular school populations. Early classification attempts used labels such as "feeble-minded," "idiot" and "moron" to identify individuals who required special programs (Hobbs, 1975a). Gradually, the terms have been replaced with labels such as "mentally retarded," "learning disabled," "mentally handicapped" and "emotionally disturbed." Although the terms seem less severe, they continue to serve as labels for groups who deviate from the norm and require special services because of their deviance. The first classes for exceptional children in the public schools were established in 1871 in New Haven (Hobbs, 1975b). These were ungraded classes which were established for children who could not be served in the regular schools. They provided placement for disadvantaged, delinquent and slow children. During the 1890s additional programs for "backward" children were established in major cities throughout the United States (Hobbs, 1975b). Children placed in these classes were described as those unable to benefit from regular school. These children were separated from the mainstream, but classification practices were not well defined. The development of intelligence testing by Alfred Binet in the early 1900s marked the beginning of the mental testing movement and provided the means for classifying individuals for educational purposes. Special education classes for the retarded with resultant segregation of these groups were considered necessary educational practice in the first

PAGE 25

18 half of the century. Lewis Terman revised the original Binet test and adapted it for use in the United States. He noted the differences in performance among children and particularly noted the poor performance of economically deprived children. He wrote (cited in Kamin, 1975): Children of this group should be segregated into special classes. They cannot master abstractions, but they can often be made efficient workers, (p. 318) However, by mid-century, the negative effects of discrimination were being recognized. The need for special educational services was acknowledged, but consideration was being given to the effects of labeling. In 1945, the Council for Exceptional Children concluded that, although special placement might be necessary in the early years, children should be returned to the regular class as soon as possible because of the detrimental effects of social isolation (Robb, 1946). The influence of special education classification on attitudes toward and interactions with children by the significant others in their lives has continued to be a topic of educational concern. Expectancy Effects Rosenthal and Jacobson's early investigation of expectancy effects and self-fulfilling prophecy was presented in their text, Pygmalion in the Classroom (1968). Their study has been severely criticized on methodological grounds (Snow, 1969), but these researchers sparked the interest of many investigators and numerous subsequent studies have been conducted in this area. Rosenthal and Jacobson (1968)

PAGE 26

19 defined self-fulfilling prophecy as "how one person's expectation for another person's behavior can quite unwittingly become a more accurate prediction simply for its having been made" (p. vii). Since this early inquiry, the expectancy phenomenon has been the subject of considerable research. In education, the primary interest in expectancy effects relates to their influence on teacher-student interaction. Carl Braun (1976) described the process as follows: The teacher perceives competencies and potentialities of children differently and these expectations are reflected in his interactions with children to produce differential performance among learners, thus fulfilling his prophecy. Teacher expectation, self-fulfilling prophecy, and teacher faith have been coined to imply this tendency for the teacher to create a reality commensurate with his perceptions. Furthermore, the learner, while creating his own reality, shadows substantially the reality forming in the teacher's mind, (p. 185) Various aspects of teacher-pupil interaction have been observed and analyzed as part of the inquiry concerning expectancy effects. Basic questions have involved evidence for the existence of expectancy effects as they relate to teacher-pupil interaction; elements which influence development of expectancies such as previous information, cummulative records, IQ, race, sex, and label; and the influence of expectancy effects on teacher and student behavior. Early research by Pippert (1969) investigated the effect of informing teachers that a group of children were considered "highly creative" on their eventual performance on a test of verbal ability. He selected twenty fourth grade children and twenty sixth grade children

PAGE 27

20 for the experimental groups. He wrote letters to the teachers at the beginning of the school year and informed them that several of the children in each group were "highly creative." These children were to be observed but were not to be encouraged to give superior performances. At midterm the teachers were sent a second letter to remind them of the creative students. At the end of the study it was found that the labeled children showed significantly better performance than the controls on a posttest measure of verbal ability. Beez (1970) conducted an experiment on preschool children who were attending a Head Start summer school. He randomly assigned thirty children to each of two groups designated as either high or low ability. The instructors were provided with background information on each child which included an IQ score and teacher observations concerning the child's school progress. Beez found that the instructors' attitudes were more favorable toward the children identified as "high ability" and that this group attained higher achievement scores at the end of the summer school session. McDonald and Elias (1976) employed a quasi experimental method to investigate the influence of teacher expectations on student achievement. They followed the progress of a group of students who showed similar performance on achievement tests at the beginning of the academic year but had been identified by teachers as having different potentials. Analysis of the end of the year achievement test results showed that students designated as those with high potential scored better than those of low potential.

PAGE 28

21 Various factors which comprise expectancy effects have been shown to include elements such as facial attractiveness, sex, race, socioeconomic status, previous school records and standardized test scores. Kehle (1974) asked teachers to rate students on personality characteristics after they had been shown a photograph and given a fictitious written description of the student. It was found that expectancy effects were evident as a result of interactions among the variables of race, sex, attractiveness and intelligence. Finn (1972) investigated teacher expectation in relation to sex, race, ability and urban or suburban environment. He defined expectations as: An expectancy, or expectation set, is a conscious or unconscious evaluation which one person forms of another, or of himself, which leads the evaluator to treat the person evaluated in such a manner as though the assessment were correct. Further, he will act in a manner consistent with the assessment, (p. 390) His definition is particularly relevant to the use of labels to describe and classify children. Finn compared the responses of fifth grade teacher volunteers from outlying suburban areas on a rating scale of student performance on an essay writing task. The teachers were furnished with fictitious information identifying the child as black or white, male or female, and of high average or low average IQ. He found there was a persistent trend for teachers from the urban schools to rate performance higher when they believed the essays were written by high ability students. He did not find this bias in the responses of suburban teachers, but explained that interfering factors such as better facilities and overall better opportunities for children as well as

PAGE 29

22 the absence of blacks in the suburban classes affected the teacher expectation level. The author summarized his study by stating "in certain settings teachers do hold differential expectations for the achievement of student groups having common non-achievement characteristics" (p. 407). Teacher response to children due to differing expectations has also been investigated. Brophy and Good (1970) used the method of dyadic classroom interaction analysis to investigate teacher communication of differential expectations to first grade children. They observed teachers as they interacted with individual students whom they had previously ranked as high or low achievers. They found that children ranked as high achievers raised their hands more frequently, initiated more contact with the teacher on work-related tasks, and received more teacher praise. When the teachers asked questions of high achievers, the children were given more encouragement to respond. They were given more clues and opportunities for second responses. The low achievers showed decreased class participation and when asked classroom questions were given less opportunity to respond. The low achievers received more criticism, and boys, in particular, received more behavioral reprimands. The authors concluded their data supported the observation that teachers communicate differential expectations to children, and these expectations function as self-fulfilling prophecies.

PAGE 30

1 23 Other researchers have also noted different teacher response to high ability students. Chaiken, Sigler and Derlega (1974) videotaped teachers in tutoring sessions with students. When working with those students who had been identified as "bright," the teachers smiled more frequently, nodded their heads, looked directly at the students and leaned closer to them. Kester & Letchworth (1972) observed teachers in classrooms after they had been given fictitious information on several class members. The teachers offered more encouragement and were friendlier toward the brighter students. Rist (1970) conducted a longitudinal study beginning with a kindergarten group. He found that the children were placed in groups during the first two weeks of school based on estimated ability levels and generally remained in these groups throughout the year. He found that children of lowest ability received less teacher contact and were placed at a table that was the most distant from the teacher. This made it difficult to hear and difficult to obtain the teacher's attention. The original group assignment of the children was maintained throughout the first and second grade. Investigations have also been made of learner response to differential treatment. Peters (1971) reported on a classroom study in which the brown-eyed children were designated as superior for a day and the blue-eyed children were considered inferior for a day. Peters described the reactions of the children as an acceptance of the label and alterations in performance and behavior which were consistent with label descriptions.

PAGE 31

24 Videbeck (1960), in an earlier study, assigned two groups of speech students to read six poems for a visiting expert. One group was praised and the other was criticized. He then assessed self-adequacy of the students and found those who had been praised improved in feelings of self-worth and those who had been criticized decreased in feelings of self-worth. Expectancy effects, though complicated and difficult to analyze, have been demonstrated to be real components of teacher-student interaction. The extension of expectancy effects from general education settings to special education programs has particular relevance to the practice of classification and labeling children. Foster, Ysseldyke & Reese (1975) conducted a study to determine the influence of expectancy effects on student teachers enrolled in a special education course which included material on teacher bias and self-fulfilling prophecy. The students were randomly assigned to an experimental or control group, and both groups were shown a videotape presentation of a normal fourth grade boy. One group was told the child was normal and the other was told he was emotionally disturbed. After watching the tape, the students were instructed to fill out a form for referral for diagnostic evaluation for the child. They were also instructed to fill out a personality questionnaire for the child. The researchers found that teachers who had been told the child was emotionally disturbed reflected this view in their completion of the referral form and the questionnaire. In spite of the fact that the teachers were observing normal behavior, they rated the child

PAGE 32

25 according to their preconceived notion of behavior characteristics of the special education label. Teachers persisted in their perceptions even though they had received course instruction on expectancy effects and teacher bias. Another study by Ysseldyke & Foster (1978) examined the effects of the labels, emotionally disturbed and learning disabled, on initial teacher bias and on opinions of regular classroom teachers when these teachers were confronted with conflicting evidence. The authors again used an initial behavior checklist. They found that teachers responded to the label and maintained their opinions regardless of contrary evidence. Weisz (1979) investigated the relationship between teacher expectations and student performance on the development of learned helplessness in children labeled as retarded. He reported that when groups of retarded and nonretarded children with equivalent mental ages were compared on task performance, the retarded group showed deficits in voluntary response initiation, deterioration in intellectual performance after awareness of failure, and emphasis on causal attributions for failure that were due to uncontrollable causes (low ability) rather than controllable causes (insufficient effort). Weisz (1981) explained that the attitude of adults toward child failure may inadvertently cause the child to develop a condition of learned helplessness. This is accomplished through the processes of two attributional errors: overextension and discounting. Overextension refers to the practice of placing more importance on a cause than it

PAGE 33

26 deserves. That is, extending its significance beyond its logical limits. For example, a teacher may accept the fact that a child cannot perform a task which is within the limits of reasonable performance for his mental age. She may attribute his failure on the task to his low ability. If the same task were presented to a normal child with the same mental age as the retarded child and the child was unable to succeed on the task, the teacher might explain his poor performance by his lack of interest, distractibility at the time or some other explanation unrelated to ability. Discounting (Kelley, 1973) is the placement of inflated significance on one factor to the exclusion of all other possible explanations. For example, if a retarded child fails at a task, the teacher or parent may believe that the task is too difficult for the child. For a nonretarded child, the adult may explain failure by a comment such as, "He didn't try hard enough." Adults foster conditions of learned helplessness by allowing children to give up on a task rather than encouraging them to persist in their effort and expecting them to succeed (Severance & Gasstrom, 1977; Weisz, 1981). Adeie Thomas (1979) related the concept of learned helplessness to children categorized as learning disabled. She found that these children generally have prolonged experience with failure, and the resulting frustrations influence their continued performance. Thomas defined learned helplessness as a "phenomenon by which an individual learns, over a series of trials, that she/he has no control over the outcome of events" (p. 209). She related this definition to the

PAGE 34

27 performance of learning disabled students and the observation that they often show poor task persistence, low frustration threshold, and a tendency to be unwilling to attempt new tasks. Seitz and Geske (1976) questioned the effect of the label "retarded" on the perceptions of mothers and teacher trainees toward the retarded child. Using randomly selected groups of mothers and students, the researchers presented them with videotapes of mother-child dyads engaged in free play. The dyads included four mother-retarded child and two mother-nonretarded child pairs. The subjects were asked to view the tapes of a mother-nonretarded child pair, mother-retarded child pair and finally, the mother-retarded child pair after labeling of the child as retarded. After viewing the tapes, the subjects were requested to fill out a behavior questionnaire. Results of the study indicated that both mothers and trainees rated the labeled retarded children as less competent and less able to obtain information than the nonretarded children. However, they also rated the retarded children as happier and more likeable than the unlabeled children. Nevertheless, the raters did not place these children high on the social distance scale; that is, they would not choose to have these children in their homes or their neighborhoods. The authors concluded that the discrepant ratings reflected an attempt to resolve cognitive dissonance within the labelers. Although they recognized the inherent limitations in the child, they resisted rejecting or not liking the child simply because he was retarded.

PAGE 35

28 Salvia, Clark and Ysseldyke (1973) questioned teacher retention of sterotypes of exceptionality in the face of normal behavior. The subjects in this study were 165 undergraduate special education and general education students. The subjects were randomly assigned to three experimental conditions, and each group was asked to complete a questionnaire on a particular child. The first group was told the child was normal, the second group was told he was mentally retarded and the third group was told he was gifted. The three groups were then shown a videotape of the purported child and again asked to rate the child. All groups viewed the same tape of a normal child. Results of the study showed that gifted children were rated more positively than normal children in areas such as attitude toward own performance and attitude toward the task. Mentally retarded children were rated less positively than normals on all areas of the checklist. The authors concluded the expectancy effects due to the applied labels were evident, but not all pervasive. Gillung and Rucker (1977) investigated the effects of behavioral descriptions and behavioral descriptions plus special education label on the expectations of teachers. The authors utilized the Rucker-Gable Educational Programming Scale (RGEPS) (Rucker & Gable, 1973) which consists of a series of behavioral descriptions and requires the teachers to choose an appropriate educational placement. The only difference in the descriptions was that one stated the special education classification and the other did not. The authors found that both regular education and special education teachers held lower expectations for children who

PAGE 36

29 were labeled than for those who were not labeled. They concluded that special education labels carry negative connotations and that teachers see children who are labeled as having relatively more severe academic and behavioral problems and requiring more intensive educational programming than children who are not labeled, Algozzine et al. (1977) investigated the effects of the special education labels of specific learning disabilities and emotional disturbance on teacher expectations of classroom performance and behavior. The researchers asked 128 teachers to review four case studies and complete a questionnaire concerning the relative acceptability of the behaviors described. It was found that teachers appeared to have preconceived notions of behaviors appropriate for learning disabled or emotionally disturbed children. Behaviors considered expected for emotionally disturbed children were not expected for learning disabled children. The authors suggested that special education labels may serve as reference points for determining tolerable behavior. That is, part of what is considered typical behavior for a particular category exists in the mind of the observer. Funding Influences The categories of exceptional children were developed in an attempt to provide appropriate help and service for children in need (Hobbs, 1975b). Special class placement was intended to offer a child the advantages of well-trained teachers, specialized curriculum and small class size. Because of the expense involved, places were limited and testing and classification processes were considered program safeguards.

PAGE 37

30 Funding practices for special education programs have strongly contributed to the labeling process. Financial support through federal and state agencies requires regulation and assurance that the money reaches the designated groups. Funds have been made available through several agencies, and two of the largest divisions include the Rehabilitation Services Administration (RSA) and the Bureau of Education for the Handicapped (BEH). Both agencies provide monies for direct service as well as sponsorship of projects for research, demonstration and training. To allocate their resources, both agencies refer to categories of target populations and these include the labels for exceptional children (Hobbs, 1975b). Perhaps the most significant federal influence on programs for exceptional children has been the passage of the Education for All Handicapped Children Act of 1975, PL 94-142. The law states, in part: A free, appropriate public education will be available for all handicapped children between the ages of three and eighteen within the State not later than September 1 , 1978, and for all handicapped children between the ages of three and twenty-one within the State not later than September 1, 1980, that with respect to handicapped children aged three to five and aged eighteen to twenty-one, inclusive, the requirements of this clause shall not be applied in any State if the application of such requirements would be inconsistent with State law or practice, or the order of any court, respecting public education within such age groups in the State (Public Law (P.L.) 94-142, 1975, Section 612 (2) (B). This law guaranteed the right of all handicapped children to a free, appropriate public education (Behr & Gallagher, 1981). The funding provided with the law is based on reported state counts of the number of handicapped children and, therefore, is bound to the

PAGE 38

31 traditional practices of testing, labeling, and placing children in categorical programs. The law assured the growth of special education programs but also further entrenched the practice of labeling. For the first time, however, this law provided for the education of exceptional children during preschool years. Labeling has been considered a necessary prerequisite to provision of special programs for exceptional children, and the danger of negative expectancy effects has been overshadowed by the advantages of special programs. But labeling of young children is complicated by variations in developmental rates, inadequacies of testing instruments, and use of categories designed for school age populations (Mercer et al., 1979). It is possible that labels applied during the preschool years are unstable and the risk of misdiagnosis and unwarranted expectancy effects may outweigh the benefits of special programs. The purpose of this study was to investigate the relative stability of special education categories applied to a preschool age group and an elementary school age group of children .

PAGE 39

CHAPTER III PROCEDURES I ntroduction According to section 228.041 (19) and 230.23 (4) (m) 4., Florida Statutes (Orange County Public Schools, Note 1): No student shall be given special instruction or services until he is properly classified as an exceptional student. The parent or guardian of an exceptional student placed or denied placement in a program of exceptional education shall be notified promptly of such placement or impending placement or denial. The term "exceptional student" means any child or youth who has been certified by a specialist qualified under regulations of the state board to examine students who may be unsuited for enrollment in a regular class of the public schools or is unable to be adequately educated in the public schools without the provision of special classes, instruction, facilities, or related services, or a combination thereof. This statute clearly specifies that children must be certified as exceptional before they are placed in special education programs. The intent of the law may be to protect children from inappropriate placement and to protect the state from unnecessary expenditure for students who do not actually require special services, but to accomplish these aims, it requires that children be categorized and labeled before being assigned to special classes. This directive applies to all children regardless of age group. To investigate the unique aspects of special education labeling practices for preschool children, a correlational study was designed which compared a sample of preschool children to a sample of elementary school children. 32

PAGE 40

33 The researcher's hypotheses questioned the stability of special education labels applied to a preschool versus an elementary school age group. Further questions were posed concerning the stability of specific classifications, EMR, TMR, or SLD, and the influence of independent variables such as sex and race on label change. Specifically, the researcher attempted to answer the following questions: 1. Were labels applied to preschool children more unstable than labels applied to elementary age children? 2. Were some classifications more unstable than others? For the purposes of this study, the classifications under consideration were educable mentally retarded, trainable mentally retarded and specific learning disabled. 3. Were the label changes which occurred within the two groups influenced by the time of initial classification, the type of label applied, individual characteristics such as the sex or race of the child and factors such as level of parent education and single or dual parent homes? 4. Were the label changes influenced by interactions among the grade group (preschool or elementary) at the time of initial testing, the type of label applied, the sex and the race of the child? Design of Research The intent of this study was to investigate classification practices in an existent school system. Because the topic dealt with human subjects and the controversial issue of diagnostic labeling prior to special education placement, it was not possible, nor advisable, to use

PAGE 41

34 an experimental design. Any, Jacobs and Razavieh (1979) stated that many variables which are of interest in educational research are not amenable to experimental methods. They explained: We cannot randomly assign children to broken or intact homes, to high or low social class, to achievementoriented or non-achievement oriented peer groups, and so forth, (p. 289) The design of this study was correlational in nature and utilized ex post facto data. Ary et al. (1979) described correlational studies as concerned with determining the extent of relationship existing between variables. They enable one to ascertain the extent to which variations in one variable are associated with variations in another, (p. 305) Ary et al. (1979) referred to the ex post facto study as one which is used when it is not possible to randomly assign subjects to groups and to directly manipulate variables. He stated: In ex post facto studies, the changes in the independent variable have already taken place and the researchers must study them in retrospect for their possible effects on an observed dependent variable, (p. 292) Kerlinger (1973) referred to the use of ex post facto methods as the basis of some of the most important educational research. He commented : It can even be said ex post facto research is more important than experimental research. This is, of course, not a methodological observation. It means, rather, that the most important social, scientific and educational research problems do not lend themselves to experimentation, (p. 391) The utilization of a correlational design as described in this study carried the limitation to external validity of inability to assign causality (Campbell & Stanley, 1963). Since the study dealt with educational

PAGE 42

35 services for actual human subjects, application of an experimental design would have required that those subjects relegated to a control group would have been denied special education services. This would have been unethical procedure and, therefore, was not implemented. Campbell and Stanley (1963) clarified the issue of causality and the usefulness of correlational techniques: Such data are relevant to causal hypotheses inasmuch as they expose them to disconfirmation . If zero correlation is obtained, the credibility of the hypothesis is lessened. If a high correlation occurs, the credibility of the hypothesis is strengthened in that it has survived a chance of disconfirmation. To put the matter another way, correlation does not necessarily indicate causation, but a causal law of the type producing mean differences in experiments does imply correlation, (p. 64) The design was limited in what Campbell and Stanley (1963) referred to as threats to internal validity by the element of selfselection. Self-selection occurs when the sample consists of individuals who are assigned to groups due to naturally occurring circumstances. This study dealt with exceptional children, a unique population which is separate from regular education and always carries the characteristic of self-selection. In addition, data used in this research were drawn from records of children referred for evaluation by their parents, pediatricians or teachers. The sample was not selected randomly from the total population of children. Data Collection The School Board of Orange County, Florida, serves the city of Orlando and its immediate environs. It includes an area of approximately 910 square miles (Florida Statistical Abstracts, 1980) and

PAGE 43

36 a population of approximately 471,000 (Bureau of Census, 1980). During the 1980-81 school year, the Orange County School Board served a population of approximately 80,000 students (Orange County Public Schools, Note 2). Of this total, 11,283 students were served by the Department of Exceptional Child Education (Orange County Public Schools, Note 3). This department provides classes for a great variety of handicapped children and includes programs for EMR, TMR, and SLD children that were considered in this study. During the 1980-81 academic year, the number of elementary school age children in each of these categories included EMR 472 TMR 97 SLD 240 In Orange County, the preschool population, that is, the age range from three to five years, is provided both a diagnostic center and self-contained special education placement for those children qualifying for special class programs. Children are referred to the diagnostic center because of problems such as suspected developmental delay, poor speech or behavior management difficulties observed by the parents, the nursery school teacher or the pediatrician. If the results of the diagnostic testing confirm the need for special placement, appropriate recommendations are made. The basis for placement recommendations may include evidence of poor progress in preacademic areas, results of intelligence testing, deficiencies in speech and

PAGE 44

37 language development, poor performance on visual motor tasks, and adaptive behavior level which is shown to be below expectations for the child's chronological age. For elementary age children, special class placement is generally accomplished subsequent to the following process: the child has difficulty in school and may be retained in a grade or be socially promoted; subsequently, the parents are informed of the child's problems and their permission to evaluate the child is solicited. When their permission is obtained, the child is referred to the school psychologist for evaluation and to the social worker for determination of a social history and assessment of adaptive behavior. If necessary, additional evaluations from the speech therapist or specific learning disabilities teacher are obtained. After testing is administered, special education placement is recommended, if appropriate. The determinants for placement include school performance, intelligence testing, ancillary specialized evaluations and measure of adaptive behavior. To investigate current classification practices, a preschool and an elementary school age special education group were assembled. To obtain the preschool group, the records of the preschool diagnostic clinic were reviewed over the four year period extending from academic year 1977-78 through 1980-81. These records have been meticulously maintained and include data on each child that has been evaluated at the center. From these records all children who were referred for reevaluation after initial classification and placement in programs for

PAGE 45

38 EMR, TMR or SLD were selected. For each child the following information was collected: sex, race, initial IQ, reevaluation IQ, initial classification, reevaluation classification, parent educational level and number of parents in the home. Parent educational level was selected as an index of socioeconomic range (SER) and was designated by the terms high, middle and low. Educational attainment of less than high school was considered low, high school attendance was considered middle and any college attendance was considered high. Years of schooling have been shown to be directly related to occupational attainment. Breedemier and Breedemier (1978) stated : If you are in the top ten percent of education receivers the chances are about thirty-eight out of a hundred that you will be in the top ten percent of occupations. Your chances of being in that top ten percent if you are in the middle group of education receivers, however, are only about seven out of a hundred; and if you are in the bottom ten percent of education receivers, they are about one out of a thousand, (p. 151) For this sample, information in this area was obtained through the computerized basic identification card which is maintained as part of each student's county record. Missing values were encountered on 43 children, and these cases were not included in the statistical analysis when parent education was entered as a variable. The total sample contained 15 children from high SER homes, 58 from middle SER homes and 182 from low SER homes. Determination of number of parents in the home was based on information available in the psychological services files and on the computerized basic identification card. If both parents were not listed

PAGE 46

39 on the ID card, the child was considered to be living in a single parent home. Children living in foster homes were placed in a separate category and were considered as missing values when parent in home was included in the analysis. This was decided because foster homes are often temporary assignments and may have changed several times during the time period considered in this study. The total sample included 106 children living in single parent homes and 166 living in dual parent homes. To obtain the comparison group of elementary age children, the following procedures were established. In Orange County, the Department of Exceptional Education has instituted staff positions referred to as exceptional education consultants. These individuals serve as administrative representatives at all special education placement staffings. Their role is to insure that proper procedures are followed and that children are placed appropriately and according to legal guidelines. These individuals maintain a written record of each child staffed into a special education program. District guidelines (Orange County Public Schools, Note 1) require standardized procedures for review and reevaluation of exceptional students: 1. Each exceptional student's progress will be reviewed at the end of the school year as part of the annual Individualized Educational Plan review. Results of this review will be documented on an Educational Planning Conference form and attached to the student's lEP. 2. A formal reevaluation of the student's placement in the exceptional education program will be conducted every three years or as conditions warrant.

PAGE 47

40 3. A review or reevaluation will also be conducted upon the request of the parent or guardian, teacher(s) or other appropriate school personnel, (p. 13) The exceptional education consultants maintain a written record of those children who are reevaluated at any time during the school year. Additionally, school psychologists provide intellectual assessment of children, and, in Orange County, they are required to maintain a record of each child that they evaluate. The comparison group of elementary age children for this study was accumulated by reviewing the records of the exceptional education consultants and the school psychologists from academic year 1977-78 through academic year 1980-81. Children initially placed in EMR, TMR or SLD programs and referred for reevaluation were included. For the SLD group, the records were reviewed for those children attending Cherokee Elementary School, a special education facility which provides self-contained classes for EH (emotionally handicapped) and SLD children. For the EMR group, self-contained classes in seven elementary schools were reviewed. For the TMR sample, the records pertaining to Magnolia Special Education Center, the Orange County facility for TMR children, were reviewed. For each of these samples, data were accumulated regarding the sex, race, initial IQ, reevaluation IQ, initial classification, reevaluation classification, parent educational level, and number of parents in the home.

PAGE 48

41 Assumptions and Limitations of Research Assumptions 1. County records were accurate and complete. 2. District guidelines were followed in regard to initial placement of students in special classes. 3. District guidelines were followed in regard to review of appropriate placement and referral for reevaluation of students in special education classes. Limitations 1. Since county records were employed, their accuracy was dependent upon careful reporting by parents and teachers. This was a potential source of error. 2. Because of the ex post facto nature of the data, missing values were encountered in some cases on the variables of parent education level and number of parents in the home. These cases were omitted in the calculations when those variables were part of the analysis. Data Analysis The raw data collected have been summarized according to age level and classification. The subjects included in the study consisted of 127 preschool children and 171 elementary age children. These were subdivided into 67 EMR, 44 TMR, and 16 SLD preschool children and 108 EMR, 25 TMR and 38 SLD elementary age children. Table I provides a summary of the proportions of males and females, blacks and whites and numbers of children per special education classification for both the preschool and the elementary school age groups.

PAGE 49

42 Table 1 Frequency Chart for Preschool and Elementary Age Groups Frequency Preschool Elementary Total Number per group 127 171 298 By Sex Males Females 78 49 110 61 188 110 By Race White Black 89 38 76 95 165 133 By Classification EMR TMR SLD 67 44 16 108 25 38 175 69 54

PAGE 50

43 Data analysis is presented in Chapter IV and the following calculations were reported. Correlations were determined between label change and each of the independent variables. Two and three way contingency tables were generated to compare the frequency of label change within the two age groups and among the three classifications: EMR, TMR and SLD. They were also used to investigate the relative influence of each of the independent variables on label change and the effects of interactions between variables on label change. The Chisquare Test of Significance was reported and the 0.05 level was set for determining significance. Finally, the data were analyzed using the multivariate technique of discriminant function. This method may be considered the same as multiple regression when the dependent variable is membership in one or the other of two groups (Kerlinger & Pedhazur, 1973; Klett & Overall, 1972). For this study the criterion or dependent variable was designated "change" or "no change" in label. The independent variables included grade group (preschool or elementary), type of initial label (EMR, TMR or SLD) and individual characteristics of sex, race, parent educational level and number of parents in the home. Information on reevaiuation label and reevaiuation IQ were utilized for comparative and descriptive purposes . Green and Tull (1970) explained that two group discriminant analysis may involve the following objectives: 1. Determine if significant differences exist between two groups.

PAGE 51

44 2. Determine which variables account most for the intergroup differences . 3. Allow the researcher to find linear combinations of the predictor variables that enable the analyst to represent the groups by maximizing among group relative to within group separation, (p. 369) Applied to this study, the purpose of discriminant function analysis was to investigate the differences between the two groups, one whose label had changed and the other whose label had not changed, and to determine those characteristics which discriminiated most effectively between the two groups. That is, were the two groups identifiable on the basis of label change, and, if so, which within group characteristics were associated with this change?

PAGE 52

CHAPTER IV ANALYSIS AND PRESENTATION OF RESULTS The data obtained for this study were analyzed through tabulation and interpretation of contingency tables, calculation of correlations between label change and the variables under consideration and application of discriminant function analysis. Since the investigation involved two groups, this method of multivariate analysis is a form of multiple regression (Kerlinger & Pedhazur, 1973). The results of the analysis were reported as discriminant function equations and as proportions of correct and incorrect classifications of cases. These values would be comparable to regression coefficients and incremental 2 increases in R reported in multiple regression. Analysis of the First Hypothesis H^l : Among children who have attended a special education class and have been referred for reevaluation , there is no significant difference in frequency of label change between those younger than six and those six years or older at the time of their initial evaluation. The research question posed by this null hypothesis was: "Were labels applied to preschool children more unstable than labels applied to elementary school children?" This question dealt with changes which occurred within the two subgroups, preschool and elementary school age children. More specific questions concerning the "nature" of change, that is, type of label change and the relationship between change and characteristics of individuals such as race and sex, were considered by subsequent hypotheses. 45

PAGE 53

46 Data used in testing the first hypothesis consisted of values obtained from the total group of 298 preschool and elementary age children contained in the study. This large group was subdivided into 127 cases who were preschool children at the time of their initial evaluation and 171 who were elementary age students at the time of their initial testing. A contingency table was computed to compare the frequency of label change within each group upon reevaluation after placement in a special education program. In the total sample of 298 children, 85 or 28.5% changed labels after reevaluation. Within the 127 preschool children, label change occurred in 59 or 46.5% of the cases under consideration. Within the elementary age group, 26 or 15.2% of the cases experienced a label change. This represented a significant 2 difference in frequency of label change within the two subgroups, X (1) = 34.91, £=0.001. For this study, probability levels of less than 0.05 were considered significant for all reported test statistics. The relationship between grade group, preschool and elementary, of the student and the frequency of label change after reevaluation was next investigated through utilization of discriminant analysis. This procedure, as a form of multiple regression, distinguishes between groups on the basis of a linear combination of the independent variables included in the analysis. In the stepwise process, an optimallyweighted sequence of variables is selected which maximally separates the groups (Lohnes & Cooley, 1971; McClave & Dietrich, 1979; Nie, Hull, Jenkins, Steinbrenner & Bent, 1975). After the optimal sequence is established, the discriminant function coefficients are applied to the

PAGE 54

47 variables and a discriminant score is derived for each subject. The subject is then assigned to a group, in this study "change" or "no change" of label, based on the probability of group membership associated with that score. The accuracy of prediction of group assignment produced by the application of the discriminant function to the variables is indicated by the percent of cases correctly classified. In this analysis, the first variable selected in the stepwise procedure was the grade group (preschool or elementary) to which the child belonged at the time of initial evaluation. Knowledge of grade group membership enabled 68.46% accuracy in prediction of label change. That is, of the 85 cases within the total sample which changed labels, 59 could be assigned to the "change" category when the only known variable was grade group (preschool or elementary) membership at the time of initial testing. Of the 213 cases which did not change labels on reevaluation , 145 could be placed in the "no change" category. The classification matrix for this portion of the discriminant analysis is presented in Table 2. The unstandardized canonical discriminant function coefficient with grade group as the independent variable and "change" or "no change" of label as the dependent variable was incorporated into the following discriminant function equation: change = 0.16 + 1.07 (grade group)

PAGE 55

48 Table 2 Discrminiant Function Matrix Label Change as a Function of Grade Group Observed Label Number of Cases Predicted Label Change No Change No Change 213 68 145 31.9% 68.1% Change 85 59 26 69.4% 30.6% Note: Total percent of grouped cases correctly classified = 68.4%, F = 39.28, df = 1,296, £ = 0.01

PAGE 56

49 For this analysis, the variables were assigned effect (contrast) coding (Kerlinger & Pedhazur, 1973). The preschool cases were coded as 1 and the elementary cases as -1. The F_ to enter at this step was 39.28 (1,296), £< 0.01. Appendix A presents the F values and other relevant statistics for the stepwise discriminant function analysis utilized for this study. The factors presented and discussed in Chapters I and II including differential rate of development, advantages of early intervention, and inadequacies of preschool testing may help to explain the higher incidence of label change within the preschool group. This study confirmed that classification change did occur substantially more frequently when a child was placed in a special program as a preschooler than when placed in a special program as an elementary student. The data analysis supported rejection of the proposed null hypothesis. Analysis of the Second Hypothesis The second hypothesis questioned the relative magnitude of label change within the total sample among the categories of EMR, TMR and SLD. The null hypothesis was stated as: H^2: There is no significant difference in frequency of label change among the three categories: educable mentally retarded, trainable mentally retarded, and specific learning disabled.

PAGE 57

50 Directionally, it was surmised that the less severe categories of EMR and SLD would show more label change than the more severe category of TMR. This was consistent with Aldrich and Holliday's (1971) observation that identification and diagnosis of mildly handicapping conditions have been less accurate than identification of more severe conditions. The two by three contingency table categorizing change in label by type of label showed that of the 175 students initially classified as EMR, 66 changed labels upon reevaluation ; of the 69 TMR students, 16 changed labels; of the 54 SLD students, 3 changed labels. The differences in frequency of label change were significant as shown by 2 the test statistic, X (2) = 22.19, £ = 0.001. A tabulation of specific initial label to final label changes is presented in Table 3. Examination of this table showed that the majority of revisions were evident within the EMR category, and these were most often from special class (EMR) to regular class placement. In six cases, students were reclassified from EMR to TMR. These changes generally occurred when the initial assignment to an EMR class had been based on borderline placement scores, and it became evident, over time, that the student required the more restricted TMR program. There were a few cases in which the student had debilitating experiences, such as recurrent seizures, and a decrease in measured IQ was noted. In 12.6% of the cases, the students moved from EMR classes to SLD, EH, or self-contained language classes. These programs do not carry the "mentally retarded" label and, in that context, may be considered a less severe placement.

PAGE 58

51 Table 3 Initial and Final Classifications within the Total Sample Final Label Initial EMR TMR SLD Regular Language EH Total Label Class Class EMR 109 6 15 38 5 2 175 62.3% 3.4% 8.6% 21 .7% 2.9% 1.1% 100.0% TMR 11 53 3 1 0 1 69 15.9% 76.8% 4 3?, ~ • ^ o 1.5% -0.0% 1 5% 1 00 . 0% SLD 3 0 51 0 0 0 54 5.6% 0.0% 94.4% 0.0% 0.0% 0.0% 100.0% Total 123 59 69 39 5 3 298 41 .2% 19.8% 23.2% 13.1% 1.7% 1.0% 100% Note: Chi-square for "change" vs "no change" of initial label for total group: (2) = 22.19, £=0.001

PAGE 59

52 Other classification changes included movement from TMR to EMR classes and, in some cases, from TMR to SLD, EH or regular class placement. The SLD category showed the least instability with only three changes occurring. Three students moved from SLD to EMR on reclassification. This study did not differentiate placement from selfcontained SLD to resource (part-time placement) SLD as a label change because, although this represented an alteration in educational approach, it did not remove the label or the connotations which the label might have carried. In the stepwise discriminant function procedure, initial label was selected as the second variable to be entered into the analysis. Since there were three possible labels, effect (contrast) coding was assigned, and two variables for initial label, IL and IL2, were entered into the analysis. The first variable showed the contrast between the EMR and SLD labels and used the coding: EMR = 1, TMR = 0, SLD = -1. The second variable showed the contrast between the TMR and SLD cases and used the coding: EMR = 0, TMR = 1, SLD = -1 (Kerlinger & Pedhazur, 1973). The discriminant function classification matrix computed after the inclusion of IL and IL2 is presented in Table 4. With the addition of these variables, the total percentage of correct classifications improved from 68.46% to 73.15%. Comparison of Table 2 with Table 4 showed that addition of initial label to the analysis added discriminative power for the selection of those cases which did not change labels. Very little difference in predictive power was evident for the cases which changed labels .

PAGE 60

53 Table 4 Discriminant Function iViatrix Label Change as a Function of Grade Group and Initial Label Observed Label Number of Cases Predicted Label Change No Change No Change 213 53 160 24.9% 75.1% Change 85 58 27 68.2% 31.8% Note: Total percent of grouped cases correctly classified F = 24.12, df = 3,294, £ = 0.01 = 73.15%,

PAGE 61

54 The unstandardized canonical discriminant function coefficients for grade group and initial label (IL and IL2) were entered into the following equation: change = -0.24 + 0.92 (grade group) + 0.97 (IL) + (-0.31) (IL2) The £ value when IL was entered into the analysis was (1,295) 27.46, significant at the 0.01 level. When I L2 was entered, an £ (1,294) of 0.18 was obtained which was not significant at the 0.01 level. The equivalent £ which represents a composite value and is a test of the significance of the combined effects of ail the previously entered variables was (3,294) 24.12, significant at 0.01. The results of the contingency table calculations and the discriminant analysis supported rejection of the second null hypothesis. They confirmed that the EMR classification was the least stable of those included in the study. Labels generally changed from more severe to less severe categories, and many changes resulted in no further need for special class services. This implied that those children had grown in independence and had shown progress in academic skills. Decreased need for special class placement is a bonus for a school system as it attempts to fit its program needs to increasingly limited budgets. These results are in partial agreement with Aldrich and Holliday's (1971) statement that severe labels are more often accurately diagnosed. That is, the TMR group showed less change than the EMR group. However, SLD is considered a relatively less severe category in regard to academic ability and IQ qualifications than EMR, and yet, very little

PAGE 62

55 label change was evident in this group. Those initially labeled as SLD tended to retain that label at least through the first reevaluation period . Analysis of the Third Hypothesis The third hypothesis questioned the "nature" of the label changes which took place. Specifically, the relationships among change, the grade group of the student at the time of initial placement, the type of initial classification, and individual characteristics of the student, such as race and sex, were examined. The hypothesis was stated: H 3: There is no significant variation in classification change attributable to the variables of sex, race, initial classification and grade group (preschool or elementary) or interactions among these variables. It was possible to accumulate information concerning educational level of the parent and number of parents in the home as described in Chapter III, and these data were included in the investigation. Information was also collected concerning initial IQ, reevaluation label and reevaluation IQ. These values were utilized for comparative and interpretive purposes. Correlations were first determined between change of label and each of the independent variables. For this portion of the analysis, dummy coding (0,1) rather than effect coding (1, 0, -1) was used. These values are presented in Table 5. The highest correlations existed between change and the grade group of the student (preschool

PAGE 63

56 Table 5 Correlations Between Change in Label and Independent Variables Variable Correlation Grade Group 0. 34 (preschool = 1, elementary = 0) p=0. 000 Initial Label (IL) 0. 24 (EMR = 1, TMR + SLD =0) p=0. 000 Initial Label (IL2) -0. 06 (TMR = ^, EMR + SLD = 0) p=0. 132 Sex 0. ,10 (male = 1 , female = 0) p=0. ,045 Race 0. .07 (white = 1 , black = 0) p=0. .102 SER 1 0. .02 (high = 1, middle + low = 0) p=0. .337 SER 2 0, .12 (middle = 1, high + low = 0) p=0, .018 Parent in Home (1 ) -0 .02 (single = 1 , both = 0) p=0 .371 Parent in Home (2) -0 .04 (both = 1 , single = 0) p=0 .273

PAGE 64

57 or elementary) and change and the initial label (EMR 1, TMR+SLD = 0). Sex, race, level of parent education (SER), and number of parents in the home showed minor correlations with label change. Their contributions were minimal in comparison to grade group and initial label. To examine the differences between the two subgroups, preschool and elementary, correlations were next determined for each subgroup separately. These results are presented in Table 6. Comparison of the two sets of values showed there was relatively greater correlation between change and each of the independent variables in the preschool group than in the elementary group. A relatively large comparative difference was seen in the correlations of 0.41, £ = 0.001 (preschool) and 0.19, £ = 0.007 (elementary) between change and initial label (IL). Tables 7 and 8 present the frequency of label changes which occurred within each grade group. The Chi-square value obtained for the preschool group was (2) 24.13, £ = 0.001, whereas for the elementary group a value of (2) 6.15, £ = 0.05 was determined. These values reflected the previously established findings that, overall, change occurred more frequently within the preschool group and specifically within the EMR classification.

PAGE 65

58 Table 6 Correlations Between Change in Label and Independent Variables for Preschool and Elementary Groups Correlation Preschool Elementary Variable Sex 0. 19 0. 04 (male = 1, female = 0) p=0. 018 p=0. 287 Race 0. 13 -0. 15 (white = 1 , black = 0) p=0. 079 p=0. 026 Initial Label 0. 41 0. 19 (EMR = 1, TMR & SLD = 0) p=0. 001 p=0. ,007 Initial Label -0. 21 -0, .08 (TMR = 1, EMR & SLD = 0) p=0. ,001 p=0, .053 SER 1 -0. .05 -0. .05 (high = 1, middle & low = 0) p=0. .273 p=0, .275 SER 2 0, .22 -0. .12 (middle = 1, high & low = 0) p=0. .006 p=0. .053 Parent in Home (1 ) -0. .002 -0. .004 (single = 1 , both = 0) p=0. .493 p=0 .480 Parent in Home (2) -0. .09 -0, .03 (both = 1, single = 0) p=0, .154 p=0 .362

PAGE 66

59 Table 7 Change in Initial Label for Preschool Subgroup Observed Label Initial Label EMR TMR SLD No Change (count) 23 30 15 (col. %) 34.3 68.2 93.8 Change (count) 44 14 1 (col. %) 65.7 31.8 6.2 Note: Chi-square for "change" vs "no change" of initial label for 2 preschool subgroup: X (2) = 24.13, £ = 0.001

PAGE 67

60 Table 8 Change in Initial Label for Elementary Subgroup Observed Label Initial Label EMR TMR SLD No Change (count) 86 23 36 (col. %) 79.6 92.0 94.7 Change (count) 22 2 2 (col. %) 20.4 8.0 5.3 Note: Chi-square for "change" vs "no change" of initial label for 2 elementary subgroup: X (2) = 6.15, £ = 0.05

PAGE 68

61 Contingency table analysis was next applied to investigate differences among other variables suggested by the comparison of the correlation tables. Interpretations in this portion of the analysis were complicated by unequal numbers and small sample sizes within subgroups. It was possible to make descriptive statements relevant to the sample under consideration, but more general applications would be unsubstantiated . Examination of label change in regard to sex showed that 60 (31.9%) out of 188 males and 25 (22.7%) out of 110 females in the total 2 group received a change in classification, X (1) = 2.87, £ = 0.09. Although the difference was not significant, it was large enough to indicate that the sex of the student was a contributing factor to label change, but its importance was minor in comparison to the more powerful influences of grade group and initial label. Overall, males tended to change labels slightly more frequently than females. When the two variables, sex and grade group, were analyzed in relationship to change, it was found that the subgroup of preschool males showed the greatest frequency of label change. Females were less numerous than males in both grade groups, but the lowest frequency of change among all subgroups occurred within the elementary age females. The tabulation of change by grade group by sex is presented in Table 9.

PAGE 69

62 Table 9 Label Changes per Grade Group, Preschool or Elementary According to Sex, Male or Female Observed Label Preschool Male Female Elementary Male Female Total Male Female No Change (count) (col.%) 36 46.2 32 65.3 92 83.6 53 86.9 128 68.1 85 77.3 Change (count) (col. %) 42 53.8 17 34.7 18 16.4 8 13.1 60 31 .9 25 22.7 Note: Chi-square for "change" vs "no change" for the total group by sex: X (1) = 2.87, £ = 0.09

PAGE 70

63 The incidence of label change was next explored in regard to race. It was found that 165 (55.4%) cases in the total sample were white and 133 (44.6%) cases were black. On reevaluation , 52 (31.5%) of the white students changed categories and 33 (25.0%) of the black students changed categories. Label change occurred slightly more frequently 2 among whites than blacks, but the shift was minimal, X (1) = 1.67, £ = 0.5. Interpretation was complicated by the distribution of race within subgroups. The preschool group was predominantly white, 89 (70%) and the elementary group predominantly black, 95 (55%). When grade group and race were considered in relationship to change, it was found that the greatest frequency of change occurred within the preschool white students. Blacks were present in proportionately greater numbers in the elementary group but changed labels more often in the preschool group. The lowest frequency of change was seen in the white elementary age students. These results are presented in Table 10. When level of parent education (SER) was considered, 15 (5.0%) cases of the total sample were in the high range, 58 (19.5%) were in the middle range and 182 (61.1%) were in the low range. There were 43 (14.4%) cases with missing values on this variable, and these were omitted from the calculations. Label change occurred in 5 (33.3%) cases in the high range, 23 (39.7%) in the middle range and 45 (24.7%) in the low range. Although change occurred most frequently in the middle range, the differences in percentages were insignificant, (2) = 5.13, £ = 0.1 .

PAGE 71

64 Table 10 Tabulation of Label Changes per Grade Group: Preschool or Elementary According to Race, White or Black Observed Label Preschool White Black Elementary White Black Total White Black No Change (count) (col. %) 44 49.4 24 63.2 69 90.8 76 80.0 113 68.5 100 75.0 Change (count) (col. %) 45 50.6 14 36.8 7 9.2 19 20 52 31.5 33 25.0 Note: Chi-square for "change" vs "no change" for the total group by race: X (1) = 1.67, £ = 0.5

PAGE 72

65 Consideration of change by grade group by SER showed that label change was seen most frequently in the preschool children at the middle SER level and the elementary age students at the low SER level. Interpretation was limited by the unequal numbers in each group and the small number of cases in the high group. The results are presented in Table 11. Number of parents in the home was next considered in regard to change of label. Change occurred in 45 (27.1%) cases of students living in single parent homes and 29 (27.4%) cases of students living in dual parent homes. There was very little difference in percentage, and, therefore, the variable had essentially no effect on label change for this sample, (1) = 3.15, £ 0.10. When change by grade group by number of parents in the home was considered, there was little differentiation ascribable to the additional variable. The percentages of change within the two grade groups remained nearly the same. These results are presented in Table 12. In an effort to more clearly identify the "nature" of the label changes which occurred, contingency tables were next computed to explore the relationship among grade group, type of label change, and the sex or race of the student. Chi-square values were not determined because, due to the increased number of cells in each cross tabulation, many cells were empty or contained small numbers.

PAGE 73

m o 4-1 _J c E o ^ 0) XI >\ S(0 4-' c 0) E fl> LU O) XI T3 D) T3 0) > 1. XI H O -I 66 00 IT) ^ ro •qCM o rr— O II QJ in o 00 O) CM 00 ro in II 00 o to in 00 to 00 00 to C7) CO 00 00 in cvj c:^ CM o CM in 'ain 00 CM in TCO in a 3 O s_ O) "(5 4-> o o "4O) c m r. (J o c in > ~(U O) c TO x: o s_ o <*(U 3 o 2

PAGE 74

67 Table 12 Tabulation of Label Changes per Grade Group, Preschool or Elementary According to Number of Parents in the Home, One or Two Observed Preschool Elementary Total Label One Two One Two One Two No Change (count) 22 43 55 78 77 121 (col. %) 53.7 57.3 84.6 85.7 72.6 72.9 Change (count) 19 32 10 13 29 45 (col. %) 46.3 42.7 15.4 14.3 27.4 27.1 Note: Chi-square for "change" vs "no change" for total group by 2 number of parents in the home: X (1) = 3.15, £40.10

PAGE 75

68 When initial label and final label were analyzed with regard to sex within each grade group, some trends were identified. These results are presented in Tables 13 and 14. Examination of these tables showed that preschool males initially classified as EMR tended to change labels more than any other subgroup. They also tended to change to a greater variety of labels than any other subgroup. When elementary age EMR males changed classifications, they moved almost exclusively from EMR to regular class. When type of label change was considered with regard to race within each grade group, it was found that the preschool, white, EMR children changed more frequently than any other subgroup. Although black EMR children showed relatively more change if first identified at the preschool level, overall they tended to be more stable than the white group. These changes are presented in Tables 15 and 16. Discriminant function analysis was utilized to examine the effects of single variables and interactions between variables on label change. When all of the single independent variables were entered into the analysis and the stepwise procedure was invoked, the variables were selected in the following sequence: grade, initial label, sex, parent educational level, race, and number of parents in the home. This resulted in an accuracy of prediction of 77.85%. As reported previously, in the discussion of the first hypothesis, when grade group (preschool or elementary) was the only variable entered, 68.46% accuracy of prediction was obtained. Additional analysis was conducted to determine the extent to which subsequent variables improved the prediction accuracy.

PAGE 76

69 Table 13 Change in Label for Preschool Group: Initial Label by Final Label by Sex Initial Label EMR TMR Final Label SLD Regular Language Class Class EH EMR Male (row %) Female (row %) TMR Male (row%) Female (row%) SLD Male (row%) Female (row %) 9 23.1 14 50.0 7 22.6 2 15.4 0 0 1 12.5 4 10.2 2 7.1 19 61.3 11 84.6 0 0 0 0 11 28.2 1 3.6 3 9.7 0 0 8 100.0 7 87.5 11 28.2 8 28.6 1 3.2 0 0 0 0 0 0 3 7.7 2 7.1 0 0 0 0 0 0 0 0 1 2.6 1 3.6 1 3.2 0 0 0 0 0 0

PAGE 77

70 Table 14 Change in Label for Elementary Group: Initial Label by Final Label By Sex Final Label Initial Label EMR TMR SLD Regular Class Language Class EH EMR Male 46 0 1 15 0 0 Crow ?,) 74.2 0 1 .6 24.2 0 0 Female 40 0 2 4 0 0 (row %) 87.0 0 4.3 8.7 0 0 TMR Male 1 15 0 0 0 0 (row %) 6.2 93.8 0 0 0 0 Female 1 8 0 0 0 0 (row %) 11.1 88.9 0 0 0 0 SLD Male 1 0 31 0 0 0 (row %) 3.1 0 96.9 0 0 0 Female 1 0 5 0 0 0 (row %) 16.7 0 83.3 0 0 0

PAGE 78

71 Table 15 Change in Label for Preschool Group: Initial Label by Final Label by Race Initial Label EMR Final Label TMR SLD Regular Class Language Class EH EMR White 15 5 12 14 3 2 (row %) 29 4 9 8 23.5 27.5 5 9 3.9 Black 8 1 0 5 2 0 (row %) 50 0 6 3 0 31.3 12 9 0 TMR White 5 (row %) 19.3 Black 4 (row %) 22.2 SLD White (row %) 0 0 Black 1 (row %) 25.0 17 65.4 13 72.2 0 0 0 0 2 7.7 1 5.6 12 100 3 75.0 1 3.8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3.8 0 0 0 0 0 0

PAGE 79

72 Table 16 Change in Label for Elementary Group: Initial Label by Final Label by Race Final Label Initial EMR TMR SLD Regular Language EH Label Class Class EMR White 22 0 2 3 0 0 (row %) 81 . 5 0 7. 4 11.1 0 0 Black 64 0 1 16 0 0 (row %) 79 0 0 1 2 19.8 0 0 TMR White 1 18 0 0 0 0 (row %) 5 3 94.7 0 0 0 0 Black 1 5 0 0 0 0 (row %) 16 7 83.3 0 0 0 0 SLD White 1 0 29 0 0 0 (row %) 3 3 0 96 7 0 0 0 Black 1 0 7 0 0 0 (row %) 12 5 0 87 5 0 0 0

PAGE 80

73 The second step in the process sought to determine the improvement in classification rate due to consideration of initial label as well as grade group. Since initial label included all three classifications, EMR, TMR, and SLD, contrast coding was assigned, and these variables were forced into the analysis at a single level as initial label (IL) and initial label two (IL2). When grade and initial label were entered, a correct classification rate of 73.15% was attained. This confirmed that both grade and initial label were important variables in the determination of label change. The classification matrix and discriminant function equation for this step were reported in the discussion of the second hypothesis. Next, to attempt to discover the specific contributions of sex and race, these were added to the analysis after grade group and initial label had been entered. This resulted in an accuracy of prediction of 74.83% which was only slightly higher than that previously obtained. Therefore, these two variables added only a small amount to the predictive power of the analysis, or perhaps the majority of their effects were accounted for when the first two variables had been entered. The classification matrix for this step is presented in Table 17. Examination of the matrix and comparison of these values with those presented in Table 4 showed that very little change in predictive ability for either the "change" or "no change" group was evident.

PAGE 81

74 Table 17 Classification Matrix Label Change as a Function of Grade Group, Initial Label, Sex and Race Observed Label Number of Cases Predicted Label Change No Change No Change 213 49 164 23.0% 77.0% Change 85 59 26 69.4% 30.6% Note: Total percent of grouped cases correctly classified = 74.83%, F = 39.28, df = 1,296, £ = 0.01

PAGE 82

75 The unstandardized canonical discriminant function equation at this step was written as change = -0.40 + 0.84 (grade group) + 1.05 (IL) + (-0.30) (IL2) + 0.38 (sex) + 0.17 (race) The F_ value at step four when the variable, sex, was entered into the analysis was (1, 293) 8.4, £<0.01. When race was entered in step five, the F value was (1, 292) 1.70, £> 0.05. The equivalent F (5, 292) was 16.90, £<0.01. Since parent education level and number of parents in the home, as demonstrated in the contigency table analysis, were found to have little influence on label change for this study, these variables were not included in the discriminant function analysis. Finally, the analysis was run with grade group, initial label and the interaction between these two variables as the only elements. This resulted in a prediction rate of 78.52% which was higher than the previously obtained rates of prediction. This indicated that the two variables, grade group and initial label, and the interactions which existed between these variables (for example, in the total sample most changes occurred in the preschool group and the largest change in all the subgroups was seen in the EMR preschool children) resulted in the best possible accuracy of prediction. It did not mean that other variables were not important factors but suggested that their influence was accounted for by consideration of grade, initial label and their interaction .

PAGE 83

76 Examination of the classification matrix at this step, presented in Table 18, showed that addition of the interaction term improved the ability to discriminate the cases belonging to the "no change" group but decreased the ability to distinguish between the cases in the "change" group . The unstandardized canonical discriminant function equation was written as change = -0.30 + 0.61 (grade group) + 1.06 (IL) + (-0.16) (IL2) + 0.58 (Int: grade group by IL) + 0.12 (Int: grade group by I L2) The F to enter for the first interaction term was (1, 293) 11.97, £ 0.05. The equivalent F (5, 292) equaled 17.35, £<0.01. Appendix B presents other relevant statistics obtained in the step backward discriminant function procedure utilized for this portion of the analysis . The information obtained from the contingency tables and the discriminant function analysis confirmed that interaction effects were present and supported rejection of the null hypothesis in regard to the interaction between grade group and initial label. However, complete interpretation of all interactions was prevented by the limitations of small number of cases and unequal sample sizes within subgroups.

PAGE 84

77 Table 18 Classification Matrix Grade Group, Initial Label and Interactions Between These Variables Observed Label Number of Cases Predicted Label Change No Change No Change 213 23 190 10.8% 89.2% Change 85 44 41 51.8% 48.2% Note: Total percent of grouped cases correctly classified = 78.5% F = 17.35, df = 5,292, £ = 0.01

PAGE 85

78 In summary, the findings from the data analysis confirmed that labels changed substantially more frequently among children initially labeled during preschool than elementary school years. Additionally, it was found that of the three labels in the study, EMR, TMR and SLD, the EMR label showed the greatest instability. Finally, it was determined that of all the subgroups examined, the preschool, EMR, white males showed more label change than any other subgroup. The implications from these findings will be discussed in Chapter V.

PAGE 86

CHAPTER V SUMMARY, MAJOR FINDINGS, DISCUSSION, SUGGESTIONS FOR FURTHER RESEARCH Summary The purpose of this study was to investigate the practice of labeling children for special education placement and to explore differences which might exist as these classification procedures were applied to preschool and elementary age students. The study was directed to the following questions: 1. Were labels applied to preschool children more unstable than labels applied to elementary age children? 2. Were some classifications more unstable than others? For the purposes of this study, the classifications under consideration were educable mentally retarded, trainable mentally retarded and specific learning disabled. 3. Were the label changes which occurred within the two groups influenced by time of initial classification, the type of label applied, individual characteristics such as the sex or race of the child, and factors such as level of parent education and single or dual parent homes? 4. Were the label changes influenced by interactions between the grade group (preschool or elementary) at the time of initial placement, the type of label applied, and the sex or race of the child? 79

PAGE 87

80 Special education programs provided by the Orange County School Board in central Florida were the site of this study. Data were obtained from county records and psychological services files concerning placement of preschool age and elementary school age children in special education programs for educable mentally retarded, trainable mentally retarded and learning disabled children. Records spanning a four year period were reviewed and data were collected in regard to initial testing and placement recommendations, reevaluation results, level of parent education and number of parents in the home. The total sample included 298 students. This large group was composed of 127 children initially placed in special education programs as preschoolers and 171 children evaluated and placed after they had begun elementary school. The data were analyzed to answer the previously stated research questions . Major Findings It was determined that children classified and placed in special education programs during the preschool years changed labels more frequently than children classified and placed during the elementary 2 school years, X (1) 34.91, £ = 0.001. Within the preschool group, 59 (46.5%) of the students changed labels upon reevaluation and within the elementary age group, 26 (15.2%) changed labels after reevaluation. Discriminant function analysis showed that consideration of grade group, preschool or elementary, as the only independent variable with "change" or "no change" of label as the dependent variable enabled 68.42% accuracy of group prediction.

PAGE 88

81 The type of label change was next examined. It was determined that, within the total sample, the EMR label changed significantly more 2 frequently than either the TMR or SLD classifications, X (2) 22.19, £ = 0.001. Of the 175 students initially labeled EMR, 66 (37.7%) changed categories upon reevaluation . Of the 69 TMR's, 16 (23.2%) changed classifications and of the 54 SLD's, 3 (5.6%) changed classifications. Application of stepwise discriminant function analysis selected grade group and initial label as the two most influential variables, and their combined effect resulted in 73.15% correct prediction of group membership . Finally, the "nature" of the label changes which occurred was explored. The importance of the variables of grade group, type of label, sex, race, parent education level and number of parents in the home was investigated in relationship to classification changes. Contingency table analysis showed that, for the total sample, the variables of sex, race, parent education level and number of parents in the home had minor influence on the incidence of label change. Overall, males tended to change categories more frequently than females, and whites changed labels more frequently than blacks. However, interpretation of interaction effects was complicated by unequal sample sizes and small numbers of cases within subgroups. In addition, the subgroups were sexually and racially identifiable with the preschool group preponderately male and white and the elementary group male and black.

PAGE 89

82 When label change within the subgroups, preschool and elementary was analyzed, it was found that the EMR category showed the most instability with the greatest percentage of change occurring within the preschool, white males. The majority of these changes were from EMR to less severe placement. Within the preschool group, the TMR label also showed relative instability. Again, the group which evidenced greatest frequency of change was white and male. The majority of changes were transfers from TMR to EMR. The interaction between grade group (preschool or elementary) and initial label, with regard to change of label, overshadowed the importance of other interactions for this sample. That is, changes occurred most frequently in the preschool group, and the majority of the changes were from EMR to a less severe category. When grade group, initial label and the interaction between these variables was included in the discriminant analysis, 78.5% correct prediction was attained . Discussion The increased incidence of label change within the preschool group may reflect difficulties encountered in testing young children. Review of the literature revealed that diagnostic testing of preschool children is limited by factors which are unique to this age group. Young children develop at different rates and a wide range of acceptable limits exists for many developmental tasks (Mercer et al., 1979). Diagnostic testing is hindered by reliance on instruments which have few items or poor standardization for lower age ranges (Divoky, 1974). Test performance is influenced by home factors, such as proper nutrition and care, and

PAGE 90

'•1 83 experiential factors, such as language stimulation and participation in preschool programs (Cataldo, 1981). Children are sometimes assessed for programs which are predicated on previous school performance. For example, preschool programs for children with learning disabilities are difficult to define because learning disabilities traditionally imply that the child has shown deficits in school subjects such as reading, writing or spelling (Keogh, 1977). A preschool child has not had the "opportunity to fail," and SLD placement is based on estimated potential rather than proven difficulty with academic tasks. The frequency of label change in the preschool group may also be related to the efficacy of early intervention programs. During the 1960s, the Westinghouse evaluation of Head Start programs cast serious doubt on the impact of early intervention attempts (Westinghouse Learning Corporation and Ohio University, 1969). Subsequent investigation (Lazar & Darlington, 1979), however, found that gains which had been claimed to disappear by second grade in children who had attended Head Start reappeared again by the fifth grade. They also found a marked decrease in the continued need for exceptional education placement for handicapped children who had participated in early intervention programs. One of the most extensive efficacy reports has been provided by the follow-up study of the Perry Preschool Project in Ypsilanti, Michigan (Schweinhart & Weikart, 1980). This project was established in 1962 for children who were socially deprived, slow learners or mildly

PAGE 91

84 retarded. The study has accumulated information on children who entered the Ypsilanti Project at age three and are now fifteen years old. Results indicate that, when compared to a control group which did not attend preschool, those children who received early intervention have maintained overall higher scholastic achievement, demonstrated higher motivation, placed higher value on schooling, been involved in less delinquent behavior, and required fewer special education assignments or support services. Placement in special education programs is accomplished categorically because of the restrictions imposed by federal and state funding requirements (Hobbs, 1975b). Before a child is provided services, he must be evaluated by qualified educational specialists and be assigned to the appropriate exceptional education category. Although these labels are applied in an effort to provide the best educational program for the child, they indicate conditions of less than normal functioning and they segregate the child from his peers. The relationship between labels and expectations of deviant behavior has been well documented (Becker, 1963; Braun, 1976). Extensive investigation has also been made of self-fulfilling prophecy (Finn, 1972; Pippert, 1969; Rosenthal & Jacobson, 1968) and particularly of the effect of teacher expectation on student behavior (Brophy & Good, 1970; Rist, 1970). Numerous studies have established the importance of expectancy effects in the use of special education labels (Algozzine et al., 1977; Mercer et al., 1979; Ysseldyke & Foster, 1978).

PAGE 92

85 Analysis of the data from the present investigation supported the literature which indicated that diagnostic evaluation of preschool children is a tenuous process. The relatively high incidence of label change may also reflect the efficacy of early intervention efforts. Regardless of the reason, the classification changes which occurred must have presented a confusing picture for parents and teachers whose expectations were influenced by the labels the children carried. Reaction of parents to identification of a handicapped child has been related to the grief process attendant to any substantial loss. Clinical studies have identified the grief process as a passage through stages of shock, despair, guilt, withdrawal, acceptance and finally adjustment (Parks, 1977). As an extension of this explanation, other researchers (Wikler, Wasow & Hatfield, 1981) have suggested that parents must go through this process over and over again as they encounter new crisis periods such as entry into school, a birthday, or any incident which again declares the "difference" or "deviancy" of their child. Rather than reaching adjustment they become involved in a life situation which the authors termed chronic grief. Bray, Coleman and Bracken (1981) investigated the significance of a variety of interpersonal, intrapersonal and situational events on families with handicapped children. Their subjects included 169 families with children with handicaps such as mental retardation, learning disabilities, emotional disturbance and physical handicaps. Their survey results indicated that the six events most significant to parents included securing appropriate educational services, receiving the initial

PAGE 93

86 diagnosis, intrapersonal family member impact, medical management, incomplete/inaccurate diagnosis and prognosis. Two of these concerns, initial diagnosis and incomplete/inaccurate diagnosis, are particularly relevant to the practice of labeling preschool children. These findings illustrate the importance of special education classifications to parents. Parents view these labels as serious pronouncements and, like a medical diagnosis, rely on their accuracy. If they are given incorrect information, parents experience needless anxiety. The presence of a handicapped child has also been shown to effect family relationships and stability. Price-Bonham and Addison (1978), in a review of the literature, found that family stress, the divorce rate, and the suicide rate were increased among parents of handicapped children. Cummings (1970) reported increased depression, lack of self-esteem and poor relationship gratification between parent and child among fathers of mentally retarded children when compared to fathers of normal children. When professionals assign labels to children which declare them as "exceptional," they automatically attach expectations concerning that child's behavior. They affect attitudes of the parents, relationships in the family and the performance of the child. Because of these extenuating circumstances, it is crucial that special education labels be used with caution. If labels applied to preschool children are relatively unstable over time, then the advantage of providing special programs may be outweighed by the disadvantages of the labeling process. It is possible that alternative placement methods, such as noncategorical programs, should be considered.

PAGE 94

87 The "Labeling Game" (Hedges, 1980) is played frequently and perhaps this is understandable under the present funding conditions. But it is a practice which is subject to error and which carries the possibility of enduring consequences. Because of the effects of labels, special educators must monitor their use and modify their application when necessary. Suggestions for Further Research The results of this study suggest the following areas for additional research : 1. This investigation was limited to one county school system. It could be replicated in counties with different demographic conditions or, on a larger scale, utilizing data from several counties or states. 2. This study did not consider other groups of exceptional children including those initially labeled emotionally handicapped, language disordered, hearing impaired, vision impaired or physically handicapped. Comparison of labeling practices among these groups might yield valuable information. 3. Although studies such as the longitudinal analysis of the Perry Preschool Program (Schweinhart & Weikart, 1980) have verified that early intervention programs are effective, a longitudinal study to determine the outcome of children served in Orange County early intervention programs has not been conducted. The present investigation has shown that classification changes occurred relatively more frequently within the preschool group, but information was not obtained on the length of time children retained their new placement.

PAGE 95

88 For example, it was not determined how long children transferred to regular class were able to keep pace with their classmates and remain in that placement. A longitudinal investigation within the Orange County School System could utilize the subjects assembled for the present study. The results of such an investigation could be extremely helpful for future program planning. 4. Questions concerning parental understanding and acceptance of labels applied to preschool children were suggested as the data were gathered for this study. Although the labels were explained to parents when their children were placed in special classes, it was difficult to determine if parents fully understood their implications. For example, could parents differentiate between terms such as educable mentally retarded and trainable mentally retarded? If their child was called learning disabled, could they comprehend the educational meaning of this term? The case files accumulated for this study could be utilized to obtain information concerning these issues. 5. The results of this study strongly suggest the need for investigation of methods of funding programs for preschool exceptional children. Rather than including this age group with all other age groups and basing allocations on student counts, alternative procedures could be developed. Possibilities might include provision of funds based on county or state populations, previously demonstrated incidence of children who required special services, or evidence of particular need for expansion of early intervention programs.

PAGE 96

89 6. If funding procedures were different, then alternative placement procedures could be considered for preschool children. Specifically, noncategorical classrooms which could provide special education services without the disadvantages associated with labeling have been suggested. Studies on the efficacy of early intervention programs have compared early education models to determine the single best approach, but it has been found that various curricula are effective. The success of programs may be due to common qualities rather than individual differences (Moore, Fredericks & Baldwin, 1981). Because of these observations, as well as the dangers associated with labeling, changes in special education delivery systems, including establishment of noncategorical resource rooms (Idol-Maestas, Lilly & Lloyd, 1981) or noncategorical self-contained classes are being implemented. Comparative studies between traditional categorical programs and the newer noncategorical approaches are needed in areas such as pupil progress, differences in expectancy effects on parents and teachers, and methods of teaching. 7. PL 94-142 carries within its provisions the mandate that exceptional children be provided an education in the least restrictive environment. Schools are attempting to fulfill this requirement by mainstreaming, that is, placing special education students in regular classes for particular subjects or for portions of the school day. At the preschool level, peer imitation and interactive play offer an opportunity for growth in understanding, development of cooperative and compassionate attitudes, and the possibility for a challenging

PAGE 97

90 environment for both the normal and the handicapped child (Cooke, Ruskus, Apolloni & Peck, 1981; Guralnick, 1981; Vincent, Brown & Getz-Sheftel , 1981.) The combination of the advantages of mainstreaming and noncategorical placement is an area which invites current research. Theoretically, it should be easier and more effective to mainstream exceptional education students who have not been assigned special education labels. Follow-up studies which compare the social adjustment and academic progress of labeled versus nonlabeled preschool exceptional education students are needed to determine the importance of and best methods of mainstreaming this age group. Conclusion Classification of children for special class placement is an organizational tool which has enabled us to fund programs and administer specialized services for students. But labels applied to human beings are not simple tags which can be assigned, reassigned, changed, or removed without consequence. Investigations are continuing on the process of labeling and the hidden meanings which labels may possess. Classification of preschool children is an area which offers more questions than answers and has many opportunities for continued research. In an article on future trends in education of preschool handicapped children, Anastasiow (1981) referred to Arthur Miller's play/ All My Sons , in which the statement was made that a person's responsibility for his acts extends beyond himself to all who are

PAGE 98

91 affected by those acts. In the play, the passage referred to a manufacturer of airplane parts. His responsibility for producing quality parts extended to all those who flew in his planes. The findings of this study demonstrated the particular relevance of these remarks to the classification of preschool children. When special educators categorize children, they must acknowledge that their responsibility extends to the child, parents, teachers and all who are affected by that label. It is hoped that the results of this study will encourage school systems to consider alternative means of funding programs for this age group and perhaps eliminate the need for labeling these young children.

PAGE 99

APPENDIX A DISCRIMINANT FUNCTION ANALYSIS STEPWISE PROCEDURE

PAGE 100

CM CO 14D CM 00 'SO o c\j 01 ^ 00 To rvj 0) D ns CM i I _i o — a •i-j If) m 00 CM CM un o o O) CM r'din a a 05 0) LD LO CD O O to O CD O LO CM 00 'flCO 00 T^ ± =! 10 q: in a t/) u c U C X! E (0 OJ £ E 3 If) c 0) _ O O O CO o 00000 00000 00000 CO rO 00 r-* 00 oo 00 r~O O O d CD Ln o 00 r>o 00 O CO 00 I^ 0000 c *-» 0) I/) T3 (DC 'D fNj X U O S, J Q) ID (J o _ q: C93

PAGE 101

APPENDIX B DISCRIMINANT FUNCTION ANALYSIS STEP BACKWARD PROCEDURE

PAGE 102

LL C 0) OJ > a CD (NJ 00 CM 05 CO CO CVJ 05 CO O I 00 '3CD in rro o cn CT) ro CO 00 o CO CM »0) n 0) X) tZ fO ru > 0 CVJ CM -I -J C C a cn CM CO XI cm"^ CM CD -"T O O CM CD in c:^ ^ CM CD CO to CD "aIo To CO o CM 0. T3 • • (D CM •!-) +-> s 1 -J C C o CO CM CM CD in in CO CM o I rin rT CD in 00 CO CD o o o CO (D o •Q . . m CM (-> 4-> i I -J c c o in Q. •!-> in a 0) in 95

PAGE 103

o o o o o o o o o o o o o o o o o o o o 00 ro r>. 00 00 CO o o o o o Tto CD CO C\J O (£) O rin VCO 0 o o o 1 I 0) T3 1. c CNJ 2 (A

PAGE 104

REFERENCE NOTES 1. Orange county public schools. Exceptional education staffing procedures . Orlando: Orange County Public Schools, 1980. 2. Orange county public schools. Student projections, 1981-82 . Orlando: Orange County Public Schools, 1981. 3. Orange county public schools. Exceptional education student count. Orlando: Orange County Public Schools, 1980. 97

PAGE 105

REFERENCES Abeson, A., & Zettel, J. The end of the quiet revolution: The education for all handicapped children act of 1975. Exceptional Children , 1977, 44, 114-130. Aldrich, R., & Holliday, A. The mental retardation service delivery system project (Research Report No. 3). Seattle: Health Resources Study Center, University of Washington, 1971. Algozzine, B., Mercer, C. D., & Countermine, T. The effects of labels and behavior on teacher expectations. Exceptional Children , 1977, 44, 131-132. Anastasiow, N. The needs of early childhood education for the handicapped: A song for the 80's. Journal of the Division for Early Childhood , 1981, 2, 1-7. Ary, D., Jacobs, L., & Razavieh, A. Introduction to research in education (2nd Ed.). New York: Holt, Rinehart and Winston, 1979. Becker, N. S. Outsiders : Studies in the sociology of deviance . New York: Free Press of Glencoe, 1963. Beez, W. V. Influence of biased psychological reports on teacher behavior and pupil performance. In M. B. Miles & W.W. Charters, Jr. (Eds.) Learning and social settings : New readings in the social psychology of education . Boston: Allyn & Bacon, 1970. Behr, S., & Gallagher, J. Alternate administrative strategies for young handicapped children: A policy analysis. Journal of the Division for Early Childhood , 1981, 2, 113-122. Braun, C. Teacher expectations: Sociopsychological dynamics. Review of Educational Research , 1976, 46, 185-213. Bray, N., Coleman, J., & Bracken, M. Critical events in parenting handicapped children. Journal of Division for Early Childhood , 1981, 3, 26-33. 98

PAGE 106

99 Breedemier, M. E., & Breedemier, H. C. Social forces in education . Sherman Oaks, California: Alfred Publishing Co., 1978. Brophy, J., & Good, T. Teachers' communication of differential expectations for children's classroom performance: some behavioral data. Journal of Educational Psychology , 1970, §2, 365-374. Bureau of Census. Preliminary reports . Washington, D. C: U.S. Dept. of Commerce, Bureau of the Census, 1980. Campbell, D., & Stanley, J. Experimental and quasi-experimental designs for research . Chicago: Rand McNally College Publishing Co., 1963. Carroll, C. F., & Reppucci, N. D. Meanings that professionals attach to labels for children. Journal of Consulting and Clinical Psychology, 1978, 46, 372-374. Cataldo, C. Early childhood assessment in schools: Its needs and strengths. School Psychology International , 1981, 2, 10-13. Chaiken, A., Sigler, E., & Derlega, V. Nonverbal mediators of teacher expectancy effects. Journal of Personality and Social Psychology , 1974, 30, 144-149. Cohen, S., Semmes, M., & Guralnick, M. Public Law 94-142 and the education of preschool handicapped children. Exceptional Children , 1979, 45, 279-285. Cole, R. & Dunn, R. A new lease on life for education of the handicapped: Ohio copes with 94-142. Phi Delta Kappan , 1977, 59, 3-10; 22. Cooke, T. P., Ruskus, J. A., Apolloni, T., & Peck, C. A. Handicapped preschool children in the mainstream: Background, outcomes, and clinical suggestions. Topics in Early Childhood Special Education , 1981, 1, 73-84. Cummings, S. The impact of the child's deficiency on the father: A study of fathers of mentally retarded and of chronocially ill children. American Journal of Orthopsychiatry , 1976, 46, 246-255. Divoky, D. Education's latest victim: The LD kid. Learning , 1974, 3, 20-25. Dunn, L. Special education for the mildly retarded--ls much of it justifiable? Exceptional Children , 1968, 34, 5-22.

PAGE 107

100 Erikson, K. Notes on the sociology of deviance. Social Problems . 1962, 9, 307-314. Erikson, K. Wayward Puritans : A study in the sociology of deviance. New York: Wiley, 1966. Finn, J. Expectations and the educational environment. Review of Educational Research , 1972, 42, 387-410. Florida Statistical Abstracts . Gainesville, Florida: University Presses of Florida, 1980. Foster, G., Schmidt, C, & Sabatino, D. Teacher expectancies and the label "learning disabilities." Journal of Learning Disabilities , 1976, 9, 58-61. Foster, G., Ysseldyke, J., & Reese, J. I wouldn't have seen it if I hadn't believed it. Exceptional Children , 1975, 41, 469-473. Gillung, T., & Rucker, C. Labels and teacher expectations. Exceptional Children , 1977, 43, 464-465. Goffman, E. Stigma . New York: Jason Aronson, 1974. Green, P., & Tuli, D. Research for marketing decisions . Englewood Cliffs, New Jersey: Prentice-Hall, 1970. Guralnick, M. The efficiency of integrating handicapped children in early education settings: Research implications. Topics in Early Childhood Special Education , 1981, 2/ 57-72. Hedges, W. In the mainstream: The labeling game. Instructor , 1980, 89, 176. Hobbs, N. The futures of children . San Francisco: Jossey-Bass, 1975. (a) Hobbs, N. (Ed.) Issues in the classification of children (2 vols.). San Francisco: Jossey-Bass, 1975. (b) Idol-Maestas, L., Lilly, M.S., & Lloyd, S. A noncategorical approach to direct service and teacher education. Exceptional Children , 1981, 48, 213-220. Kamin, L. Social and legal consequences of IQ tests as classification instruments: Some warnings from our past. Journal of School Psychology , 1975, 13, 317-323.

PAGE 108

101 Kehle, T. J. Teacher's expectations: Ratings of student performance as biased by student characteristics. The Journal of Experimental Education , 1974, 43, 54-60. Kelley, H. The process of causal attribution. American Psychologist , 1973, 28, 107-128. Keogh, B. Early ID: Selective perception of perceptive selection? Academic Therapy , 1977, 12, 267-273. Kerlinger, F. Foundations of Behavioral Research (2nd ed.). New York: Holt, Rinehart and Winston, 1973. Kerlinger, F., & Pedhazur, E. Multiple regression in behavioral research. New York: Holt, Rinehart and Winston, 1973. Kester, S. W., & Letchworth, G. A. Communication of teacher expectation and their effects on achievements and attitudes of secondary school students. Journal of Educational Research , 1972, 66, 51-55. Kitsuse, J. I., & Spector, M. Toward a sociology of social problems: Social conditions, value judgements and social problems. Social Problems , 1973, 20, 407-419. Klett, J., & Overall, J. Applied Multivariate Analysis . New York: McGraw-Hill, 1972. Lazar, I., & Darlington, R. Summary report, lasting effects after preschool . (DHEW Publication No. (OHDS) 79-30179.) Washington, D. C.: U.S. Government Printing Office, 1979. Lessen, E., & Rose, T. State definitions of preschool handicapped populations. Exceptional Children , 1980, 46, 467-469. Lohnes, W. , & Cooley, P. Multivariate data analysis . New York: Wiley & Sons, 1971 . MacMillan, D., Jones, R., & Aloia, C. The mentally retarded label: A theoretical analysis and review of the research. American Journal of Mental Deficiency , 1974, 79, 241-261. Matza, D. Becoming deviant . Englewood Cliffs, New Jersey: Prentice-Hall, 1969. McClave, J., & Dietrich, F. Statistics . San Francisco: Dellen Publishing Company, 1979. i

PAGE 109

102 McDonald, F., & Elias, T. Beginning teacher evaluation study : Phase II technical summary, final report . Princeton: Educational Testing Services, 1976. Mercer, C. Algozzine, B., & Trifiletti, J. Early identification: An analysis of the research. Learning Disabilities Quarterly , 1979, 2, 12-24. Mercer, J. The struggle for children's rights: Critical juncture for school psychology. School Psychology Digest , 1977, 6, 4-19. Moore, M. G., Fredericks, H. D., & Baldwin, V. L. The long-range effects of early childhood education on a trainable mentally retarded population. Journal of the Division for Early Childhood , 1981, 4, 94-110. Nie, N., Hull, C, Jenkins, J., Steinbrenner , K., & Bent, D. Statistical package for the social sciences . New York: McGraw-Hill, 1975. Parks, R. Parental reaction to the birth of a handicapped child. Health and Social Work , 1977, 2, 52-66. Peters, W. A class divided . New York: Doubleday and Co., 1971. Pippert, R. A study of creativity and faith. Manitoba Department of Youth and Education Monograph , 1969, 4. Powell, T. Educating all disabled children: A practical guide to PL 94-142. Exceptional Parent , 1978, 8, L-3; L-6. Price-Bonham, S., & Addison, S. Families and mentally retarded children: Emphasis on the father. The Family Coordinator , 1978, 3, 221-230. Public Law 94-142: Education for all handicapped children act of 1975. Statutes at large, 89, 1975. Public Law 94-142: What does it mean to you? The Bridge , 1981, 2/ Random House Dictionary of the English Language . New York: Random House, 1966. Rist, R. Student social class and teacher expectations: The self-fulfilling prophecy in ghetto education. Harvard Educational Review, 1970, 40, 411-451.

PAGE 110

103 Robb, I. Segregation versus nonsegregation of exceptional children A report of panel discussions at the twenty-second annual meeting of the international council for exceptional children. Exceptional Children , 1946, 12, 235-240. Rosenthal, R., & Jacobson, L. Pygmalion in the classroom . New York: Holt, Rinehart and Winston, 1968. Rucker, C, & Gable, R. The Rucker-Gable educational programming manual . Storrs, Ct. : Rucker-Gable Assoc., 1973. Salvia, J., Clark, G., & Ysseldyke, J. Teacher retention of sterotypes of exceptionality. Exceptional Children , 1973, 39 , 651-655. Schweinhart, L., & Weikart, D. Young children grow up: The effects of the Perry Preschool Program on youths through age 15. Ypsilanti, Mich.: The High Scope Press, 1980. Seitz, S., & Geske, D. Mothers' and graduate trainees' judgment of children: Some effects of labeling. American Journal of Mental Deficiency , 1976, 81, 362-370. Severance, L., & Gasstrom, L. Effects of the label "mentally retarded" on casusal explanations for success and failure outcomes. American Journal of Mental Deficiency , 1977, §2, 547-555. Shanker, A. Public Law 94-142: Prospects and Problems. The Exceptional Parent , 1980, 10, 51-56. Shur, E. M. Labeling deviant behavior . New York: Harper & Row, 1971. Snow, R. Unfinished pygmalion. Contemporary Psychology , 1969, 14, 197-199. Terman, L. M. Feeble-minded children in the public schools of California: The menace of feeble-mindedness. School and Society , 1917, 5, 161-165. Thomas, A. Learned helplessness and expectancy factors: Implications of research in learning disabilities. Review of Education Research , 1979, 49, 208-226. Tuckman, S. The placement of pseudo-retarded children in classes for mentally retarded. Academic Therapy , 1972, 7, 165-170. Videbeck, R. Self-conception and the reaction of others. Sociometry , 1960, 23, 351-362.

PAGE 111

104 Vincent, L. J., Brown, L., & Getz-Sheftel , M. Integrating handicapped and typical children during the preschool years: The definition of best educational practice. Topics In Early Childhood Special Education , 1981, I, 17-24. Weisz, J. R. Perceived control and learned helplessness in mentally retarded and nonreretarded children: A developmental analysis. Developmental Psychology , 1979, 15, 311-319. Weisz, J. R. Effects of the "mentally retarded" label on adult judgments about child failure. Journal of Abnormal Psychology , 1981, 90, 371-374. W estern States Technical Assistance Resource (WESTAR). Monmouth, Oregon: author, 1981. Westinghouse Learning Corporation and Ohio University. The impact of Head Start experience on children's cognitive and affective development . Springfield, Virginia: U.S. Department of Commerce Clearinghouse, 1969. Wikler, L., Wasow, M., & Hatfield, E. Chronic sorrow revisited: Parent versus professional depiction of retarded children. American Journal of Orthopsychiatry , 1981, Sl^, 63-70. Ysseldyke, J. E., & Foster, G.G. Bias in teachers' observations of emotionally disturbed and learning disabled children. Exceptional Children, 1978, 44, 613-615.

PAGE 112

BIOGRAPHICAL SKETCH Patricia Maher was born in Van Wert, Ohio, in 1942. Her family moved to Florida in 1946 where she received her elementary and secondary education. She graduated from Bishop Moore High School, valedictorian, in 1960. She received her Bachelor of Science degree in pre-medicine, biology, graduating "With Distinction" from Rollins College, Winter Park, Florida. She completed a year's internship in medical technology and worked for three years as a medical technologist in hospitals within the Orlando area. In 1969 she attended the University of Virginia in Charlottesville and obtained a Master of Education degree in education of emotionally disturbed children. She returned to Orlando and has been employed by the Orange County School Board since 1970. She has worked with various classifications of exceptional children including emotionally handicapped, mentally retarded, and learning disabled. During this time she also completed courses at the University of South Florida, Florida State University and the University of Central Florida to obtain certification in school psychology. She has served as the psychologist for the preschool diagnostic center in Orange County, Florida, for the past six years. She is currently completing the doctoral program in curriculum and instruction, instructional leadership, at the University of Florida. 105

PAGE 113

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Arthur J. Lewisd Chairman Professor, Instnactional Leadership and Support I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Barry J;;' 0uinagh Associate' Professor, Foundations of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Robert S. Soar Professor, Foundations of Education This dissertation was submitted to the Graduate Faculty of the Division of Curriculum and Instruction in the College of Education and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May, 1982 Dean for Graduate Studies and Research



PAGE 1

COGNITIVE-BEHAVIOR MODIFICATION: THE USE OF SELF-INSTRUCTION STRATEGIES BY FIRST GRADERS ON ACADEMIC TASKS By CHRISTOPHER WILLIAMS SPARKS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1986

PAGE 2

Copyright 1986 by Christopher Williams Sparks

PAGE 3

To Judd, Sara, and Steve with thanks and love

PAGE 4

ACKNOWLEDGEMENTS For both the product and the process of my time as a doctoral student, I offer sincere thanks to My committee, Gordon Greenwood, John Newell, Linda Crocker, and Suzanne Krogh, whose guidance and example have introduced me to worlds of information and new parts of myself; My teachers, whose gifts of challenge and knowledge have been invaluable; My family, for signs of pride and words of encouragement when tasks seemed overwhelming; and My good, good friends, for the big favors, little presents, and constant faith they extended over what, amazingly, turned out to be years. iv i

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS iv ABSTRACT vii CHAPTERS I INTRODUCTION 1 Statement of the Problem 1 Self-Instruction Training 2 Self-Instruction Research 4 The Purpose of the Study 6 Significance of the Study 10 Limitations of the Study 11 Hypothesis 11 II REVIEW OF THE LITERATURE 13 Historical Development 13 The Shift from Operant Conditioning ... 13 The Impact of Language Development Studies 15 Influence of Verbal Mediation and Task Analysis 16 Relevant Metacognition Research 17 The Formal Self-Instruction Model 19 General Studies in Self-Instruction 20 Meichenbaum ' s Original Study 20 Self-Instruction Studies Since 1971 ... 21 Academic Studies in Self-Instruction 23 Guidelines for Effective Research in Self-Instruction 28 Summary of the Literature Review ' 29 III METHODOLOGY 31 Research Questions and Hypothesis 31 Method of the Study 32 Subjects 32 Instruments 33 Procedure 43 Statistical Design 52 V

PAGE 6

Page IV RESULTS AND CONCLUSIONS 54 V SUMMARY AND RECOMMENDATIONS 6 2 Summary 62 Conclusions 64 Implications 65 Recommendations 67 APPENDICES A INVITATION TO PARTICIPATE 69 B SELF-INSTRUCTION TRAINING 71 C GUIDELINES FOR EFFECTIVE SELF-INSTRUCTION. . . 74 D MEICHENBAUM' S SELF-INSTRUCTION MODEL 76 E INSTRUCTIONS FOR THE EXPERIMENTAL GROUP. ... 77 F INSTRUCTIONS FOR THE CONTROL GROUP 82 G RESEARCH DESIGN 84 REFERENCES 85 BIOGRAPHICAL SKETCH 88 vi

PAGE 7

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COGNITIVE-BEHAVIOR MODIFICATION: THE USE OF SELFINSTRUCTION STRATEGIES BY FIRST GRADERS ON ACADEMIC TASKS ^ by \ Christopher Williams Sparks May 1986 Chairman: Gordon E. Greenwood Major Department: Foundations of Education The purpose of this study was to apply the cognitivebehavior modification self-instruction model, which involves the use of cognitive modeling and children's private speech to develop children's self-guiding statements and images in task performance, to actual classroom practice and content. Fifty-six first grade students were randomly assigned within intact reading groups to two treatment conditions. Students in the experimental group were trained by their teachers in self-instruction on non-reading tasks during four sessions separate from reading instruction. Both experimental and control groups participated in regular, basal-directed reading instruction. The dependent measure for this study was children's performance on six selected independent seatwork tasks provided by the reading series. vii

PAGE 8

A multivariate analysis of variance (MANOVA) using a two-way nested design was used to test the hypothesis that there were no significant differences between experimental and control groups on their overall task performance. Univariate analyses of variance (ANOVAs) were then used to determine if the overall difference between groups existed for each task. The MANOVA analysis yielded a highly significant difference between groups, followed by the ANOVAs' smaller but significant differences on five of the six tasks. From the above results, the following conclusions were drawn. First, children who used self -instruction applied academic skills with significantly higher accuracy through increased skill in problem definition, strategy selection, self-evaluation, self-reinforcement, and error coping. Second, self -instruction , as a teachingthinking process which focuses on the processes students use in task performance, is generalizable across at least some tasks. Third, self-instruction can be effectively used to teach planful behavior in academics with groups of normal children in classroom settings. Fourth, teachers can be easily and effectively trained to incorporate self-instruction into their curricula. Fifth, cognitive strategy changes can be demonstrated on ordinary performance in naturalistic settings. All conclusions point to promising gains through further applied cognitive-behavior modification research. viii

PAGE 9

CHAPTER I INTRODUCTION Statement of the Problem One of the major challenges in education today is the development of students' independent thinking skills and work habits. Children often acquire information in a passive, rote fashion, and then are unable to respond to questions and solve problems unless those questions and problems closely resemble the learning task. This sort of "patterned learning" effect has been particularly apparent to this researcher (a teacher) in school-based reviews of standardized test errors and math problemsolving mistakes, in children's composition efforts, and, in the middle school years, their inability to work on class or home assignments without guidance or supervision. One approach to increasing children's personal control over learned information is to train them in selfinstruction, a cognitive control technique developed within the cognitive-behavior modification paradigm by Donald Meichenbaum and others in the early 1970s. By the end of the 1950s operant conditioning had begun to evolve from various forms of laboratory -1-

PAGE 10

-2stimulus-response investigations to behavior therapy programs with children. Educators in the 1960s became familiar with behavioral objectives, behavior modification, token economies, and precision teaching. "Modeling," "shaping," and "positive reinforcement" became part of everyday professional language and the theme of teacher inservice programs in the early 1970s. Further, during the 1970s behavior therapy underwent a "shift from an emphasis on modification of attentive and disruptive motor behaviors to a concern with educational tasks that involve cognitive or thinking skills" (Meyers & Craighead, 1984, p. 4). This transition was fostered by the work of social learning theorists which provided cognitive explanations of modeling effects, and by the recognition that behavior therapy, or simple operant conditioning, did not produce generalizable or long-lasting behavior change. Selfinstruction evolved from cognitive-behavior modification investigations of cognitive processes and strategies children use in task performance. Self-Instruction Training In self-instruction training, cognitive modeling and private speech are combined as an approach to teaching thinking skills. Children are taught how to use selfstatements and images to think and plan behavior; they are i

PAGE 11

-3not instructed in what to think or encouraged to focus on "right answers." Although right answers would appear to be a logically correct educational goal, both metacognitive and self-instruction researchers reject that aim and set the development of underlying thinking skills or executive processes as a priority of teaching efforts. Self-instruction in task performance requires that the student engage in problem definition, appropriate strategy selection, self-evaluation of performance, self-reinforcement, and coping behaviors. To date, the self-instruction process has been used with young children on social, interactive, motor, and academic tasks. Meichenbaum' s model includes the following steps: 1. cognitive modeling (an adult performs the task and talks to self) ; 2. child performs the task with model's direction for guidance (overt external guidance) ; • 3. child performs task and talks aloud to self (overt self-guidance); 4. child whispers and performs task (faded, overt self-guidance) ; 5child performs task via inaudible or private speech or nonverbal self-instruction (covert self-instruction). (Meichenbaum, 1977, p. 32) At step three the child's skills in problem definition, planful behavior, and implementation of strategies can be observed.

PAGE 12

-4Self-Instruction Research The cognitive-behavior modification investigations utilizing self-instruction during the 1970s and early 1980s have demonstrated success in modifying subjects' performance. Research designs have included, usually, single subjects of specialized (hyperactive, learning disabled) populations in laboratory settings. Tasks presented have been sensorimotor or interpersonal, and treatment effects have been measured on tests other than direct performance (such as Porteus Mazes, standardized tests, or the Matching Familiar Figures Test) . These designs have been apropos for theoretical development and, in effect, piloting the use of a new intervention. Such intervention research is normally prompted by some need in a clinical population and focuses on resolving various dysfunctions for a long period before being used in a preventative or developmental fashion. If self-instruction is to be proved a useful way to enhance the development of thinking skills in normal early learners, previous research designs must be altered. To begin with, utility with groups of normal subjects on academic tasks in naturalistic settings must be demonstrated. Once such success is established, Craighead, Meyers, Craighead, and McHale (1983) indicate that client assessment, generalization of treatment effects, and loci of

PAGE 13

-5interventions, and use of paraprof essionals will become the issues of the 1980s. In other words, the testing of the self-instruction model as a teaching-thinking program in classroom settings can begin to address the question of its practicality. Meichenbaum sets the stage for selfinstruction research in academics by stating. We can conceive of academic tasks where the teacher provided the children with a set of tasks . . . and the child's job was to identify what the problem is, how he or she will go about solving the task, where the likely pitfalls are, etc. . . . Teachers could give assignments and ask the children to describe in detail how they are going to go about performing the assignment. . . . Discussion could center on the process, not only the product, of the assignment. (Meichenbaum, 1985, p. 421) A number of other studies have investigated selfinstruction in designs which would not be replicable or appropriate in regular classrooms (see Meichenbaum, 1985, for a review of these studies) . There have been only two doctoral studies done (Rhodes, 1979; Sullivan, 1981) which have begun to approach the full applied Meichenbaum model. In one of these studies (Sullivan, 1981), "selfinstruction" was interpreted as verbalization of a given problem in mathematics rather than as the development of self-guiding strategy statements and was used in conjunction with a mathematics problemsolving strategy, yielding no information about the metacognitive advantages of self-instruction per se. In the other study (Rhodes, I

PAGE 14

-61979) , the self-instruction process was abbreviated until only its cognitive-modeling component was taught, and, again, the effects of the full self-instruction model were not assessed. The Meichenbaum model does involve cognitive modeling and overtto-covert skills practice, but these procedures must further include opportunity for the child to engage in problem definition, strategy selection and implementation, and error management. This study provides necessary information regarding the feasibility of utilizing the full self-instruction model in classroom situations, with normal children, on academic tasks. The Purpose of the Study The purpose of this study is to apply the cognitivebehavior modification self-instruction model as developed by Meichenbaum to actual classroom practice and content. The design includes the use of the full self-instruction model without abbreviation, extension, or collaboration with other problemsolving methods and is in direct response to the following needs statements in general reviews of the field. 1. Kendall (1977) , in a discussion of obtaining generalization of learning through use of cognitivebehavior modification techniques, stated "the focus of the training materials and the setting contribute

PAGE 15

-7meaningf ully to the type of generalization achieved" (p. 331) , and suggested that the way to promote "the natural development of verbal mediation" (p. 331) and to "foster each child's personal cognitive control" (p. 331) in the classroom is to use psychoeducational materials in a teacherstudent situation. 2. Meichenbaum and Asarnow (1979) maintained that self-instruction training should teach "cognitive skills or executive routines that are transsituational" (p. 28) and that such training should "focus directly and explicitly on the skills and tasks that are to be learned, not on some presumed underlying deficit" (p. 30). 3. Hobbs, Moguin, Tyroler, and Lahey (1980) questioned the demonstrated utility of cognitive-behavior modification with children, indicating that the common problems with this research area were the specificity of the independent variable, lack of results from natural settings and normal performance, and lack of information on the impact of treatment in classroom or home situations. 4. Meichenbaum and Burland (1981), commenting on the self-instruction format, stated, "At this point we can share the general training strategy that is being employed and make a call for more research to assess the pedagogical potential of cognitive-behavior modification procedures" (p. 112).

PAGE 16

-85. Kazdin (1982) emphasized the need to show that "changes in cognitive skills are reflected in measures of ordinary performance in naturalistic situations" (p. 77 ) and stated that when the target population is "not identified on the basis of deficits of dysfunction, the criteria for significant change may be demonstration that the intervention produces a significant increment above the previously accepted normative level" (p. 78 ) . 6. Meichenbaum (1985) pointedly stated the need for classroom teacher participation in self-instruction research and set guidelines for the development of teachingthinking classes to be integrated into the school curricula. In this study, first grade teachers from among volunteers in a local school district were randomly assigned to treatment or control groups; selfinstruction supplemented regular reading instruction for the treatment groups. Teachers in the treatment conditions were trained in the principles and administration of self-instruction, and they trained their students in self-instruction. The investigation took place in normal first grade classroom settings; academic tasks from the established curriculum were presented to the students in both treatment and control groups to be done independently. For the purpose of this study, the following terms are defined by specific characteristics.

PAGE 17

-9"Normal first graders" are those children in regular classrooms who have not been retained in the grade and who are not staffed into any special program. It is assumed that students not identified or referred for special education programs are "normal" or "average" in that they possess a range of abilities exclusive of the very high and the very low. A "classroom setting" refers to a nonlaboratory educational situation consisting of one teacher and approximately 25 students. The students involved in the study were not participating in any other school-based programs (tutoring, precision teaching, etc.). "Academic work" means written basal reading assignments in the district-adopted reading series, Ginn. These workbook pages provide a set of work for the children which has content validity and direct relevance to skills taught. Children are tested for correct placement in this series. "Self-instruction training" is the use of images and self-guiding statements to direct behavior. "Private speech" refers to egocentric or self-addressed verbalizations; it is not intended for a listener to hear or to communicate with others.

PAGE 18

-10Significance of the Study The original self -instruction studies targeted hyperactive /impulsive children in an effort to teach those children to engage in planful behavior, to be aware of and use cognitive strategies as skills increased in selfcontrol or interpersonal behavior. There is still interest in changing the reflective/impulsive balance or cognitive tempo of such specialized populations, but applied research in self -instruction is needed to extend knowledge of its effectiveness on academic tasks and its use as a preventative or developmental approach to problems apparent in specialized populations. In normal populations, using selfinstruction to teach students planful behavior in academics remains to be investigated and may provide a remedy for those problems which are easily visible in school settings: students' inadequate study skills, inability to work independently, and inability to solve problems in any fashion other than that patterned from the teacher. The positive results of this study will substantially extend the limited existing information in this area and will contribute to the literature on the explicit training of metacognitive strategies.

PAGE 19

-11Limitations of the Study The following limitations must be considered in this study. 1. Teachers were selected from a volunteer pool recruited by open invitation across the district. These teachers' professional curiosity and willingness to participate in experimentation may be indicative of other qualities which make them "more effective" or "more dynamic" than others. 2. The treatment condition could have been misinterpreted or incorrectly administered by individual teachers in spite of careful training; the control groups could have been affected by their teachers' needs for them to do well. Hypothesis Treatment teachers participating in this study supplemented regular basal-directed reading instruction with students' training in self-instruction. Measures of normal classroom performance were taken on six sets of basal reading tasks completed independently by the children. The control groups received regular basal-directed reading instruction and completed the same assignments as the treatment groups, but without selfinstruction training.

PAGE 20

-12The supposition was that children with selfinstruction training would apply academic skills with significantly higher accuracy because of increased skills in problem definition, strategy selection and implementation, and error management. The multivariate null hypothesis developed was as follows: There will be no significant differences between the experimental and control groups in terms of performance on a set of six separate academic tasks including: word identification in context and reality/fantasy discrimination; decoding verbs with inflected _s; word insertions in sentences and puzzles; sequencing pictures or sentences; completing sentences with inflected-s_ verbs; advanced word insertions in sentences and in puzzles.

PAGE 21

CHAPTER II REVIEW OF THE LITERATURE Historical Development The Shift from Operant Conditioning Cognitive-behavior modification experimentation began in an attempt to advance results gained in operant conditioning: That is, the intent was to develop the subject's capacity to retain skills or patterns learned and to generalize those skills and patterns across different situations and tasks. Such retention and transfer were not generally a product of operant conditioning. Craighead (1982) identifies three major factors leading to the shift, in the 1970s, from operant to cognitive-behavioral interventions with children. The first was a cognitive information processing explanation of modeling effects; the second was development of selfcontrol interventions; and the third, independent of the first two, was developments in cognitive therapy. Cognitive information processing . In the area of modeling or observational learning, Craighead (1982) cites Bandura's (1969) work on the roles of attention and -13-

PAGE 22

-14retention as the beginning of new interpretations of behavior therapy procedures. Bandura's social learning theory proposed that cognitive and environmental influences are mediated by the cognitive processes of perception and attribution, and that the person and the environment influence one another. From the tenets of social learning theory sprang laboratory-based investigations of children's self-mediational strategies in enhancing self-control on such tasks as delay of gratification and resistance to temptation (see Meichenbaum, 1985, for a review of these studies). Self-control . Self-control studies, which had been placed within the realm of operant conditioning for the duration of behaviorism's influential decade, began, in the 1970s, to be conceptualized as more cognitive in nature. Craighead (1982) cites Kanfer's work dividing self-control into the components of self -monitoring, self-evaluation, and self-reinforcement as advances in clinical application and research in self-control interventions. These "advances" later contributed significantly to the principles involved in Meichenbaum' s selfinstructional format. Cognitive therapy . Cognitive-behavior therapy with adults began in attempts to "combine the clinical concerns of cognitive-semantic therapists with the behavior therapy technology" (Meichenbaum & Burland, 1979, p. 425).

PAGE 23

-15Craighead (1982) maintains that fundamental to cognitive therapy is the assumption that maladaptive cognitions such as inappropriate, irrational and illogical self-statements (based on personal assumptions and beliefs) produce psychological disorders which are best alleviated by modification of those cognitions. Cognitive-behavior therapy with children was given impetus as this relationship between cognitive and behavioral/physiological events began to be established and continued to develop as a therapeutic alternative to psychoanalysis or behavior therapy in the treatment of childhood disorders, Meichenbaum ' s use of self-statements in behavior change evolved independently of this work in cognitive-behavior therapy, but against its favorable backdrop. The Impact of Language Development Studies Soviet research . Added to the converging influences of social learning theory, self-control interventions, and cognitive-behavior therapy was the work of Soviet psychologists in language development. According to Vygotsky (1962), children's speech progresses from an obvious externally oriented, overt stage, through an egocentric, self-addressi ve phase, to inner speech. Vygotsky suggests that inner speech is a foundation for children's thinking. The overt-to-covert

PAGE 24

-16language development model suggested to Meichenbaum that "private speech is initially facilitative and then drops out of the repertoire with the development of task proficiency" (Meichenbaum, 1977, p. 23). Meichenbaum' s use of private speech . In 1969 Meichenbaum' s doctoral research involved training schizophrenics to emit "healthy talk" to control their verbalizations and behavior; his success in this area led him to ask if other clinical populations could be explicitly trained to selfinstruct and spontaneously generate statements to guide behavior (Meichenbaum, 1977) . Influence of Verbal Mediation and Task Analysis Verbal mediation . Coincidental to the developments in cognitive-behavior modification were efforts being made in the analysis of "verbal mediation deficiencies. " In the area of verbal mediation, the use of task-appropriate mediators is seen as involving comprehension of task, production of strategies, and, importantly, implementation of those strategies (Meichenbaum & Asarnow, 1979). While the cognitive-behavioral selfinstruction process does not call for such emphasis on a phase-division of mediational intervention with children, it does call for a "cognitivefunctional analysis of children's task performance" (Meichenbaum & Asarnow, 1979, p. 12).

PAGE 25

-17Task analysis . The cognitive-functional analysis of task performance requires an awareness of the child's self-statements and images (which can be observed during the overt self-guidance step of self-instruction) and an understanding of the task components. Meichenbaum used the work of Gagne (1964) as a model for identifying the hierarchy of behaviors within a task, but translated each step into self-statements and strategies that could be modeled by the trainer and rehearsed by the child (Meichenbaum & Asarnow, 1979). Relevant Metacognition Research The field of cognitive-behavior modification progressed parallel to and convergent with the field of metacognition. Each is concerned with cognitive strategy deficits leading to poor performance on tasks, and each is concerned with the learner's ability to identify the problem at hand, select and implement strategies, and evaluate performance. Cognition refers directly to such acts of thinking as attending, remembering, and choosing; metacognition refers to one's personal awareness of those acts. Researchers laboring under the rubric of "metacognition" tend to demonstrate interest in identifying the cognitive processes leading to deficit performance;

PAGE 26

-18researchers working in cognitive-behavior modification tend to maintain concern for the subject's performance when applying appropriate cognitive strategies. Each group is concerned with effective use of cognitive strategies and the subject's awareness of personal control over those strategies. Those of the metacognitive ilk focus more on breaking the cognitive processes into components, while the cognitive-behavior modifiers focus more on breaking the task into components. Work on metacognitive development suggests that "children fail to consider their behavior against sensible criteria, they follow instructions blindly, and they are deficient in self-questioning skills that would enable them to determine these inadequacies" (Brown, 1980, p. 457). Self-instruction (saying guiding statements to oneself) is one technique from the cognitive-behavior modification camp which is "often used to teach a person to monitor his progress, compare what he/she is going to what he/she should be doing, and self-reinforce" (Craighead et al, , 1978, p. 164); in other words, self-instruction provides a mechanism for one to engage in metacognitive acts before, during, and after task performance. In a review of studies on the use of cognitive instruction, Meichenbaum (1985) points out that children who receive cognitive instruction on specific skills without attention or training given to the development of

PAGE 27

-19superordinate processes are not able to generalize learned skills to other tasks or maintain improvements across time. On the other hand, children taught to define problems, engage in planful behavior, and monitor that behavior achieve substantial transfer of learning. The selfinstruction sequence fosters the development of those planful behaviors or thinking skills. The Formal Self-Instruction Model The original study . The research in language development and his own successful clinical use of private speech implied to Meichenbaum that a therapeutic package could be developed which would utilize self-instructional statements in behavior change; with Goodman, (Meichenbaum & Goodman, 1971), Meichenbaum formalized a self-instruction model for use with children to control impulsivity and motor behavior and to develop self-controlThe final structure of the model reflected current progress in social learning theory, self-control interventions, cognitive therapy, language development, verbal mediation, and task analysis. Current reviews . After reviewing the results of almost a decade of self-instruction investigations, Meichenbaum and Asarnow (1979) theorized that the overt speech in self-instruction training organizes the

PAGE 28

-20information available, facilitates problem definition, and aids in strategy selection and implementation. Further, they suggested that orientation and attention to task may be enhanced, and that coping skills and self-reinforcement skills become a useful part of the child's repertoire. The self-instruction format provides a content-free question set which the child can utilize across a variety of tasks. The goal of cognitive-behavior self-instruction is to teach children to generate, choose, and apply cognitive strategies, not to increase a specific skillsbase. Meichenbaum' s (1985) review of the field indicates that the time has come to investigate the use of selfinstruction in classrooms on academic tasks. General Studies in Self-Instruction Meichenbaum' s Original Study Research in cognitive-behavior modification utilizing self-instruction began with Meichenbaum and Goodman's (1971) study in which hyperactive, impulsive preschoolers were trained to talk to themselves as they did a variety of sensorimotor (copying patterns, coloring figures) and problem-solving (completing pictorial series, solving mazes) tasks. Results of this study indicated that hyperactive, impulsive children could be taught to think before they acted and not be subject to the dominant motor

PAGE 29

-21response (Meichenbaum & Goodman, 1971) . The children trained in self-instruction had significantly better performance on a variety of standard psychological tests and demonstrated increased reflectivity on the Matching Familiar Figures Test. Further, more than half of the trained impulsive children continued to use self-instruction strategies on the posttest and in later sessions (Meichenbaum & Burland, 1981) . The Meichenbaum and Goodman study used (a) single subjects of (b) a special population in (c) a laboratory setting; (d) the tasks involved were sensorimotor and (e) were not tested on a direct performance measure, but on tests quite removed from the task. Most of the relevant investigations done in the decade following this 1971 study were different in theme but almost identical in composition; these studies are presented here to show, for historical purposes, the types of studies done in selfinstruction . Self-Instruction Studies Since 1971 The most comprehensive review of studies in selfcontrol is Meichenbaum ' s (1985) "Teaching Thinking: A Cognitive-Behavioral Perspective, " a chapter in Chipman, Segal and Glaser's Thinking and Learning Skills (Volume 2) ; Research and Open Questions, which was sent to the

PAGE 30

-22investigator by Meichenbaum when a request for material relevant to applying self-instruction to academics was made. This review includes most of the studies specified below and represents the main body of research in selfinstruction; most investigations cross-reference others which pre-date them, and they represent the range of topics studied before an interest in academic application was developed. In a review of cognitive-behavior modification with children, Meichenbaum and Burland (1981) cite its successful use to establish "inner speech control" over the disruptive behavior of hyperactive children and aggressive children; with cheating behavior; with Maze performance of hyperactive boys; over the conceptual tempo of emotionally disturbed children; and over the conceptual tempo of normal children. Other investigations cited by Meichenbaum and Burland combined self-instruction training with imagery practice and with operant procedures. Also cited are work with learning disability children; work with hyperactive children; work with retarded children; and articles on social reasoning and problem-solving using self-instruction. In another review paper, Meichenbaum and Asarnow (1979) reference work with hyperactive children on arithmetic tasks, work on verbal control tasks, and use of self-instruction on .kindergarteners ' handwriting skills.

PAGE 31

-23In an article by Hobbs, Moguin, Tyroler, and Lahey (1980), work with hyperactive children's conceptual tempo is cited. Again, the interest of this study is in the use of selfinstruction with normal children, in nonlaboratory settings, on academic work normally distributed to the children; the preceding references have been presented as sources for the interested reader and to illustrate the nature of investigative interest to this point. Academic Studies in Self-Instruction An extensive review of the literature including the Current Index to Journals in Education (CUE), Resources in Education (RIE) , Educational Resources Information Center (ERIC) , and the Dissertation Abstracts International yielded two studies concerning academic tasks and normal children in conjunction with self-instruction. These two studies are the most directly relevant done to date since they involved academic research in self-instruction with regular classroom populations. They are described below. Sullivan, 1981 . At the University of Oregon, Sullivan (1981) did a doctoral study called "A Comparison Between Attack Strategy Training and Attack Strategy Training in Combination with Self-Instruction in Teaching Academic Tasks to First Graders." Although Sullivan's investigation

PAGE 32

-24was directed at comparing results for impulsive and reflective children, and his results were measured on a researcher-developed math criterion test, various standard intelligence tests and achievement tests, all of which would have negated interest by this researcher, there were other disengaging differences illustrative of the ways in which self-instruction research is interpreted. Sullivan used direct instruction to teach math computation skills to one group, and, for the second group, coupled that instruction with self-instruction in the sense that children were "verbalizing key phrases aloud as they worked through the task" (Sullivan, 1981, p. 44). Sullivan's math problems were "academically relevant materials" (p. 8) , but were not a subsection of the students' normal math curriculum. Sullivan's study "did not assess the effectiveness of verbal self-instruction as a treatment without specific attack strategies" (p. 84), In other words, children verbalized statements of particular strategies which were taught to them; they did not generate their own self-statements and strategies. This approach fits more closely with studies concerned with the influence of self-verbalization on academic achievement (Grimm, Bijou, & Parsons, 1973; Lovitt & Curtis, 1968); these studies found that simple verbalization of problems in the math area enhanced children's performance. The true selfinstruction procedure is content-free: It calls for

PAGE 33

-25problem definition by the student, selection of strategies by the student, and self-monitored implementation of those strategies. Sullivan's " selfinstruction" group did outperform his attack-strategy-only group, and his study was based on Meichenbaum' s work, but his interpretation of the self-instruction purpose and format and its application are not in alignment with that of this researcher or with the stated guidelines for research by Meichenbaum (see next section) . Rhodes, 1979 . The University of Louisville dissertation by Rhodes (1979) is problematic in some of the same ways as Sullivan's. The author's intent was to demonstrate that Meichenbaum' s self-instruction could be used to direct readers' attention to task and to teach students to generate questions regarding the purpose and meaning of assigned reading passages. Although the results of this dissertation indicate that behavioral methods do increase reading comprehension (on standardized tests after five weeks of intervention) , the Meichenbaum model was substantially abbreviated, and the task so narrow that it was not clear that cognitive self-instruction was responsible for the increase in comprehension, or that, as a modus operandi, the children learned a taskgeneralizable self-instruction model; they may have learned simply to question the title and paragraph of any story.

PAGE 34

-26In the Rhodes study the " selfinstruction" format involved, first, the teacher's modeling or scanning the assigned story, reading the title and first paragraph, and questioning for purpose and meaning as cued by the teacher's manual; this took up one week. For the second week, students individually scanned the material, generated questions regarding purpose and meaning, and read silently. The teacher selected certain of the student-generated questions for discussion at the conclusion of reading. During the third, fourth, and last weeks these procedures were repeated, but the children's spoken questions faded to silence. In the Rhodes study children had to generate questions regarding story purpose or meaning from the title and first paragraph; there was no problem definition or strategy selection in that task, nor was there any of the necessary self-reinforcement or error management modeling. Further, it is not clear whether the control group was directed to scan and then re-read the story (without the modeling and question-generating time) , in which case the number of times each story was read and the following question-discussion could have increased comprehension for the experimental group. It appears that the Rhodes study is one in which the cognitive modeling portion of the full self-instruction model was used as a teaching strategy for a particular

PAGE 35

-27reading task; the procedures used are, actually, more akin to the SQ3R (Survey, Question, Read, Recite, and Review) method cited in Pressley and Levin's (1983) Cognitive Strategy Research; Educational Applications and in much of the reading literature. In this process, the reader scans titles, subheadings, and summaries to get an idea about the story before reading and then restates any boldface headings in question form. Again, as with Sullivan's study, in the Rhodes study there is both a distortion of the Meichenbaum model and an overlap with research in the content area, or, at best, use of a simple faded cognitive-modeling teaching strategy. Summary of academic studies . The two dissertations reviewed above present varying points of view as to the uses and interpretations of self-instruction. They do not, however, clearly apply the Meichenbaiam selfinstruction format to academic instruction. Each of these authors seemed to perceive selfinstruction as a new mode of instruction for the purpose of getting more "right answers" in a particular subject area rather than as an approach to teaching thinking skills. In order to further explore and understand the use of self-instruction with young children, the research must focus on development of strategies which are contentfree and not patterned as a "set" of attack plans for a particular "set" of problems. The dissertations reviewed

PAGE 36

-28do not address the research need or adhere to the guidelines for self-instruction research. Guidelines for Effective Research in Self-Instruction Meichenbaum (1985) makes recommendations in two areas for those who wish to do research in self-instruction training: First, he describes a point of view to which the teacher must be committed in order to effectively transmit the training, and, second, gives guidelines for the training to foster generalization. Ideally, the teacher should model a metacognitive perspective, one which indicates openness to thinking and anticipation of mistakes which, when they occur, are viewed as problems to be solved. The instructor should be careful of attributions made when failures occur, with statements directed at nurturing the problemsolving attitude. Lastly, the teacher must conduct a task analysis of skills to be taught, as selfinstructions are comprised of superordinate cognitive strategies which are implemented to successfully apply content skills. To maximize generalization, training in self-instruction must be on a task which has the same requisites or elements of the performance task-, but is overtly different from that performance task. The pupil must be a collaborator in generating the strategies to be used, and the private

PAGE 37

-29speech trained must be developed with and compatible with the child's natural style. The cognitive strategy training should take place after component skills and experience with the task are in the child's repertoire, and the pupil must be helped to recognize the new task as one facilitated by self-instruction and requiring transfer of those strategies. The instructor needs to ensure against rote repetition of patterned strategies, encouraging personal involvement by the student through use of faded cues and acceptance of the child's wording. The teacher should directly encourage the child to generalize strategies to various tasks or situations, perhaps by engaging in discussion of other suitable uses. Lastly, a sense of selfsatisfaction should be nurtured in the child, with attention given to developing the child's self-reinforcing and coping skills and flexibility in the use of strategies is fostered. Summary of the Literature Review From the foregoing review of the literature the following conclusions are apparent. 1. Self-instruction provides a mechanism for one to engage in metacognitive acts before, during, and after task performance; in other words, self -instruction can be used to teach "planful" behavior.

PAGE 38

-302. Since the original Meichenbaum study dealing with hyperactive /impulsive children on sensorimotor tasks, many investigations using selfinstruction with specialized populations on sensorimotor or interpersonal tasks have been done. 3. Most self-instruction studies have involved single subjects, special populations, and laboratory settings. Most tasks studied have been sensorimotor and have not been tested through measures of normal performance. 4. The two studies done on the use of self-instruction in academics have not interpreted self-instruction in a metacognitive sense and have not met the current needs for applied research. 5. No studies have been done utilizing selfinstruction with normal children, in a classroom setting, on normal performance of academic work.

PAGE 39

CHAPTER III METHODOLOGY Research Question and Hypothesis The following research question emerged from the review of research on the use of self-instruction with young children: Can the performance of normal first graders on basal seatwork tasks be improved through the use of self-instruction? Or, in other words, can cognitive strategy changes made be reflected to a measurably significant degree on classwork? The following null hypothesis was developed to be tested by this study: There are no significant differences between the experimental and control groups in terms of performance on any of six separate academic tasks including: word identification in context and reality/fantasy discrimination; decoding verbs with inflected s^; word insertions in sentences and in puzzles; sequencing pictures or sentences; completing sentences with inflected-^ verbs; advanced word insertions in sentences and in puzzles. -31-

PAGE 40

-32Method of the Study Subjects First grade students were chosen as the target group for this study for two reasons: the speech characteristics of their age and their academic skills level. In the state of Florida first grade students must be six years old on or before September first of the entry year. Between ages five and seven children's speech is more overt than covert; before age five speech is largely reflective and modeled from others and after age seven is likely to be internalized, or covert. The overt-speech characteristic is necessary for observation of the selfinstruction development. Self-instruction training provides the student or client with strategies for problem identification, strategy selection, and implementation. The student or client utilizes selfinstruction in conjunction with his or her skills in the problem or content area. First graders who would be considered "normal" have an adequate but not sophisticated skills-base in reading and math; such a skills base becomes the point of departure for selfinstruction training. In the reading area assigned level chosen for this study, children do recognize letters and sounds and are aware of what reading is, but have not

PAGE 41

-33begun to generate their own word attack strategies or to develop sight vocabularies. Nine teachers with a total of 56 student subjects from 2 rural and 3 city schools were recruited by open invitation to all first graders in the Alachua County, Florida, schools. Each teacher's instructional assignment for the year included a class of approximately 25 students with a range of ability in the reading area. Within each of the nine classes, the group of students in the beginning first grade instructional level was randomly assigned to the treatment or control group for the study. No school contained both treatment and control groups; each treatment condition contained both urban and rural schools. Instruments Any academic task for which instruction and standardized seatwork materials are provided would lend itself to self-instruction. For this study reading seatwork tasks were chosen because, in the Ginn district-adopted reading series, such lessons are prescribed in the teachers' manuals, pacing is set for each book, and workbook materials (the "Skillpack" and "Studybook") correlate with the instruction and pacing. These seatwork materials are intended to be presented within the regular classroom setting to be done independently by the children. Further,

PAGE 42

-34for this series children are tested for correct placement before beginning a level, ensuring that all subjects belonged at the given level of instruction. The choice of Ginn Level 3, Fish and Not Fish (Clymer, Venezky, Johnson, & Pearson, 1984) was made for several reasons. First, Levels 1 and 2 are primarily letter/sound and experiential materials without basalstructured worksheets involving reading on the students' part. Many first graders enter with sufficient skills to waive Levels 1 and 2, since those levels simply review and reinforce the kindergarten program. Level 3 is an acceptable firstinstruction book for first graders. Levels 4, 5, and 6, the remaining books for instruction in first grade, would also be appropriate for use with selfinstruction, but students' placement in those levels in the early part of the year would indicate somewhat sophisticated reading skill. Successful training might be attributed to the learner's earlier readiness for assignment to that level. For Levels 4, 5, or 6 to be tested, the timing of the investigation for normal first graders should be OctoberNovember or later. Levels 3, 4, and 5 of the Ginn reading series, all being potential choices for the investigation (depending on the implementation dates of the study) , were subjected to a task analysis by the investigator. This means that within each level specific tasks were identified for which

PAGE 43

-35cognitive strategies would be useful. These tasks included phonics exercises, word-identification in context, use of multiple meanings, identifying the main idea, sequencing, and distinguishing reality /fantasy. A set of tasks for each book was assembled from among the tasks which had a large enough item bank (9-27 items) to be useful. Level 3 thus became the appropriate selection for self-instruction training because it is normally assigned to children in first grade between ages five and seven, it has a standardized independent-work program, it is an acceptable first book for first graders, and its materials contain a sufficient item bank of specified skills to make measurement feasible; these advantages dovetailed with the timing of the study. The specific tasks from Level 3 which were involved with this study included phonics exercises, word identification in context, and sequencing, all with different underlying reasoning required of the student. The tasks, which were divided into six sittings, are described below and illustrated by Figures 3-1 through 3-6. Set 1 is word identification in context and reality/ fantasy discrimination, demonstrated by matching pictures to sentences using lines, within Unit 1 of the book (Figure 3-1) .

PAGE 44

-36Read eocn sentence. Then draw a line to the picture it moicnes. " i) 1^ The fish is tor The doiphin,The fish is tor the cot. "Look ct the fishl". "Look at the dolphir*!" "We con get fish for the doipMn.* "We con get fish to eol." The men con get the fish. The dolphm con get the ftsh. Rtad ncn senttnc*. Th»n oniw o m *o Tf ptctur* It motcncs. -Som con. i Sofo con not. Ana con. Ano con not. Jim con. Jim con not. Kv\ con. Ktn con not. ReaO eocti sentenca. Then Braw o line ro the picture it nvalches. 1Where does Bern eoi?\^;^^ Where floes the col eol? 3. The CO! will gel woter. Ano will gel water. VyU Where does Soro like lo ploy? ^
PAGE 45

-37Set 2 deals with word identification in context demonstrated by matching sentences to pictures using circles, lines, or bubbles, within Unit 2 of the book; and decoding verbs with inflected s (Figure 3-2) . Set 3 involves word insertions using sight vocabulary with key words at the top of the page, within Unit 1, and word insertions in crossword puzzles (Figure 3-3). Set 4 is sequencing pictures or sentences, within Unit 2, by assigning a number to the pictures or sentences (Figure 3-4). Set 5 involves word insertions in sentences using verbs with s inflection, with the verbs written above or below the sentence, within Unit 2 (Figure 3-5). Set 6 deals with word insertions using sight vocabulary with words at top of the page, within Unit 2, and word insertions in crossword puzzles (Figure 3-6). The performance score is the total right/wrong measure for each task set, representing the students' abilities to apply academic skills in reading to independent seatwork tasks, with or without self-instruction. Two self-instruction pilot studies were conducted. The first, in April, 1984, was done to test the success of cognitive modeling and self-guiding statements with little children. Kindergarteners preparing for the Metropolitan Readiness Test were trained in selfinstruction and found, through informal observation, to be responsive

PAGE 46

-38Figure 3-2. Performance Tasks, Set 2 (Clymer et al., 1984, used with permission.)

PAGE 47

-39-

PAGE 48

-40(IM SIMM UWtm rMKRAM »r 1 villi p«rs»**lM, lo) Reod »!• stories. Then wilie 1 nexl to me lirtt thing ihol happens in eocn sloty. Wnle 2 nexl to the next thing jhot happens. . , 1. Granaira win mix the ctoy. Then Grandma will boke it. Then Grandma wiB take K. 1 Grandma will mix the cloy, 2. Sara ond Ken took in. Sora ond Ken see Mom with the shorlt. Sara and Ken look la Sofo and Ken see Mom with the short 47) Look Qi the pfeiures in each row. Write 1 k> the picture that comes first. Wnte 2 In the picture that comes next. Write 3 In the piciure thoi comes lost. Figure 3-4. Performance Tasks, Set 4 (Clymer et al., 1984, used with permission.)

PAGE 49

-41Writs o word to finish each sentttnca. get gets Sora . TTit « ri.^. ^ _ , Read the sentSfKtt. RD In th* dni% rml lo th» word thot goes into th« sent«ncft. 1. Grandma can bread. • fTXJke O runs O 2. A fish water. O gats O likes O 3. The cot here. O loolc O ploys O puts 4. Anoond Beth . . In the von. O get O hops O loolcs 5. Beth for Jim. O coma O looks O ploys 6. Soro with the cloy. O ploy O works O gst rrM ltii#ytaa« for L»nl ). T: ilit r. ^J). af in* SIM REAAIS riKlCKU ^ T^Msr* Clvwr •*« M 3 / A I Beth 0IOV5 wiin the ball. 3ein and Jim play with tha ball. work worlts Mm M Oac«An« we* froSt<.. «l— •»* . •tU. p«r«taal«a. ' '"^ _. * ^ r^mt ^ • *^ — — ^ 1. Ken : '.i : : o le ih« von. eats gets looks IT^Jl^HX =. m3\ 2. The man a book tor K«a. ^^^kp^ looks pioYS wortcs ^^^'fC^^ 3. Ken or the book. needs bakes sees ^\ ^^5^^ 4. He a shork In the book. ^..i^^mokes puts likes 5. Ken Ihe bookl S4 Oacooinq: *a>M win uiiMciion -| Figure 3-5. Performance Tasks, Set 5 (Clymer et al., 1984, used with permission.)

PAGE 50

-42bread has dough Do 1 and 3 Ilk* this 1 Then do 2 like ihis -* . 2. You baks this. Then it is bread. 1. You can eat this. 3. Jim breed. La Figure 3-6. Performance Tasks, (Clymer et al., 1984, used with Set 6 permission . )

PAGE 51

-43to its structure and format. In May, 1985, a second pilot study was conducted with 10 first grade children to test the organization and choice of materials for this study; the materials were found to be practical and appropriate for use with self -instruction . Procedure Subject recruitment . During the preplanning session for the 1985-86 school year, curriculum resource teachers from every elementary school were addressed at their first county-wide meeting and asked to forward to the researcher a report on the number of first graders at their schools and the reading levels to which those children were assigned. Additionally, names of first grade teachers were obtained. Immediately following this report, first grade teachers at each school were contacted by the researcher and invited to participate in the study. This invitation (see Appendix A) included information to the teacher that participation would be relevant to the current emphasis on problemsolving skills development in Alachua County. Nine teachers with a total of 56 students in the beginning reading level of the Ginn series volunteered to participate in the study. Each student/teacher group was randomly assigned to treatment or control groups.

PAGE 52

-44Subject information collected for analysis included the race, sex, socioeconomic status, and age of each student. Students in the study were not in any special education or remedial education programs or participating in any other educational/tutorial or research programs. The subject recruitment was successful in creating comparable groups of normal first graders ready to be taught to read and perform independently on written tasks. Teacher training . Training sessions of approximately two hours were scheduled at the convenience of the participating teachers; experimental and control teachers were trained separately. The training for the experiment teachers included (a) information about self-instruction training or the metacognitive approach (see Appendix B) ; (b) guidelines for effective self-instruction (see Appendix C) ; (c) training on student self-instruction training tasks (described below) ; and (d) all materials for data collection (the training tasks and/or performance tasks) . Complete teacher instructions for the experimental group are in Appendix E; for the control group, instructions are shown in Appendix F. Teachers were also provided with a timetable for action which correlated with the Ginn reading pacing. The training tasks . Self-instruction training tasks should be similar to the performance tasks in terms of their need for strategy selection and implementation, but

PAGE 53

-45should not overtly resemble the performance task. For this reason, non-reading tasks requiring strategy selection and reasoning were selected to train the children in the self-questioning, self-instruction process. These tasks involved matching similar pictures, picture association tasks, completing pathways, tracing figures to completion, reasoning tasks with matching, figure completion, and recognition of quantity. The self-instruction training was done in four separate sessions, and the tasks are illustrated in Figures 3-7 through 3-11. The experimental groups teachers were trained to use overt-to-covert cognitive modeling and children's speech to formulate four self-guiding questions which included the following steps derived from Meichenbaum ' s protocol as described in the Introduction and presented in Appendix D: 1. "What should I do here?" 2. "Am I doing what I said I would do?" 3. "What if I make a mistake, how will I fix it?" and 4. "How did I do?" or "Let me check my paper before I give it to the teacher. " The children developed and practiced the self-guiding questions on the training tasks, which were broken into four sets of approximately 15-minute sessions. Thus, the experimental group children had about one hour, over four different days, of practice in self-instruction alone.

PAGE 54

Figure 3-7. Training Set 1

PAGE 55

-47Figure 3-8. Training Set 2

PAGE 56

Figure 3-9. Training Set 3

PAGE 57

-49Figure 3-10. Training Set 4

PAGE 58

-50The children were taught to apply the self-guiding sequence as they worked and to repeat the sequence over again if mistakes were found. These self-guiding questions are, by design and purpose, content-free and later aided the children in selecting strategies from their reading skills repertoires to complete the performance tasks. Teachers guided the children to develop four questions which reflected the mandated steps in the selfinstruction model, but understood the importance of accepting the children's wording of those questions. Each experimental group, with students working together, arrived at four questions which were acceptable to the children and met the format. The performance tasks . The performance tasks selected were the students' independent worksheets from the Ginn skillpack and studybook, as specified in the Instr\aments section. These follow prescribed lessons which the teachers were encouraged to follow as exactly as possible for the duration of the study. Specifically, for each student, each teacher in the experimental group was provided with (a) training tasks for self-instruction and (b) a packet of chosen reading tasks keyed to the lesson and point in the teaching series when those tasks should be distributed to the students. The control group teachers received the same materials exclusive of the training tasks.

PAGE 59

-51Implementation . Ginn Level 3 has a five-week instructional plan. Teachers in both the experimental and control groups instructed reading as usual according to the Ginn teachers' manual. Each group received the same page assignments at approximately the same time. During the third week of instruction in reading, but separately from the reading group setting, the treatment teachers provided four sessions (about 15 minutes each) of selfinstruction training using the provided training tasks. During Level 3 instructional weeks four and five, five minutes of self-instruction review took place in the reading group, and then children were given the performance tasks and data collection began. Again, the self-instruction question set was structured from the previously agreed-upon wording or private speech of each group of children and was content-f ree ; no reference was made to the work at hand, but strategies for doing work were reinforced. The control group participated in reading instruction and received the performance tasks as the experiment group did. All assignments were collected and scored by a master key by the researcher. All teachers, experimental and control, were in weekly contact with the investigator during the experimental period, and each received a visit during the study. The investigator was easily accessible by telephone or visitation to answer questions and/or model the self-instruction process.

PAGE 60

-52Statistical Design Student performance on the Ginn reading tasks, testing the hypothesis that there were no significant differences between experimental and control groups on the reading performance tasks, was analyzed using a multivariate analysis of variance (MANOVA) on a two-way nested design. As shown in Figure 3-11, teachers are nested within group and students are nested within teacher. An alternative design was also considered in which school nested within group was a factor and teachers were nested within school. As subsequent analysis revealed virtually no difference in results for these two designs, the school factor was deleted from the model in the interest of parsimony and ease of reporting results; Figure 3-11 depicts the simpler design for which results are reported in Chapter IV. Figure 3-11. Nesting Design

PAGE 61

-53A repeated measures design was considered, but since interest was in comparing overall group performance across tasks rather than group performance by (between) individual tasks, a MANOVA was used. I

PAGE 62

CHAPTER IV RESULTS AND CONCLUSIONS The purpose of this study was to compare the performance of first graders with and without self-instruction strategy training, on academic tasks. Children in the experimental (self-instruction) and control (no selfinstruction) groups each completed six sets of basal seatwork tasks distributed by their teachers to be done independently. Each task set contained a particular set of items from among the basal skills presented for study. A multivariate analysis of variance (MANOVA) on a two-way nested design was used to analyze the difference between the two groups for the six tasks. Teachers were nested within group, and students were nested within teachers. Following the MANOVA, six univariate analyses of variance (ANOVAs) were used to compare the difference between the experimental and control group on each separate task. The results of these analyses are presented below. Table 4-1 presents the means and standard deviations of the different groups on each task. -54-

PAGE 63

-55Table 4-1 Means and Standard Deviations for Experimental (N = 29) and Control (N = 27) Groups Performance Task Group Number #Items ^^21 Experimental Control 1 21 o o . oo M 20. 55 1 0 . 0 O SD 0.90 3.65 2 15 a 1 M 14.79 1 4 nn J. 4 . u u SD 0. 67 1.41 3 14 .77 M 13. 44 11.25 SD 1. 18 2.61 4 10 . 36 M 9.62 8. 00 SD 0.62 1.17 5 16 . 58 M 15.10 12. 33 SD 1.01 1. 92 6 9 .69 M 8.96 7.85 SD 0.18 1. 51 Table 4-2 shows the partial correlation coefficients between tasks 1-6 after controlling for experimental condition. Some tasks have strong correlation and others have little or no correlation; for example, the highest partial correlation observed occurred for tasks 1 and 2; the lowest occurred for tasks 2 and 4. The MANOVA was performed using SAS and Wilks' Criterion was used to analyze the overall difference between groups. This multivariate test showed a substantial difference between groups with F = 191.11 with

PAGE 64

-56VD I iH U) y. m (0 Eh cn -P C 0) •H u •H m (1) 0 U C 0 •rH -P (0 a> o CM u 1 rH •H +J xt u (13 to Eh in a; w Eh EH ro w as Eh (M W (0 EH cn Eh ro in Ph Q ro o O 00 o <^ m 00 CN o CvJ o 00 ro o in o o o o o o o o o o o o o o r-H O CO in o in ro o o CN CN O 00 o o o o CTi O o in CM O o o o o o o rH O o o I o o
PAGE 65

-576 numerator degrees of freedom and 2 denominator degrees of freedom, and p = .0052. Using a pre-established alpha level of .05, this led to a general rejection of the multivariate null hypothesis (at alpha < .05): There is a difference between the experimental and control groups in terms of performance on the six separate seatwork tasks selected from the basal reader workbooks. As a follow-up procedure univariate ANOVAs were used to determine if the overall difference between groups existed for each task. Table 4-3 shows the Fs and probabilities for each task. Table 4-3 Results of Univariate Tests by Task df Type III SS (numerator) df Type III SS (denominator) F P Task 1 1 42. 15 7 46. 59 6.33* .04 Task 2 1 7. 56 7 7.29 7. 26* .03 Task 3 1 56. 74 7 32. 94 12.06* . 01 Task 4 1 36. 72 7 7. 13 36.01* .0005 Task 5 1 106. 11 7 14. 73 50.42* . 0002 Task 6 1 11. 90 7 19. 31 4. 31 ,07 Significant for alpha < .05.

PAGE 66

-58The six individual Ftests using TEACHER (GROUP) mean square error in the denominator (the nesting effect for teacher within group) showed significant treatment effect on each of tasks 1-5, with respective p-values of .04, .03, .01, .0005, and .0002. Although the nesting model was used, the actual effect of nesting was nonsignificant; therefore, these results are conservative. For task 6, significant nesting effect was found (p = .008). The F-value for GROUP using TEACHER (GROUP) in the denominator is only marginally lower than the critical F-value required for statistical significance, with p = .07. To summarize, the effects on five of the six univariate tests are significant. While it is ordinarily good statistical practice to require a higher degree of significance in repeated testing to protect against compounded Type I errors, two aspects of the present study give added protection against compounded Type I error. First , reported univariate F-values used TEACHER (GROUP) as the denominator, even though the nesting effect was not significant. F-tests using residual error as the denominator produced p-values of .017, .016, .0004, .0001, and .0001. The reported F-values are therefore taking a conservative approach and retaining the integrity of the nested model. Second , the MANOVA F was very highly significant (p = .0052), which gives global protection against compounded Type I error of repeated univariate testing.

PAGE 67

-59To summarize, first graders trained in selfinstruction performed significantly better on basal seatwork tasks presented to them to be done independently. The first task called for word identification in context and reality/fantasy discrimination. This task required matching sentences to pictures and involved reasoning skills. Task 1 was done better by the selfinstruction group. The second task also called for word identification in context, but with a different format: Children had to circle pictures at the end of each sentence, and there were "distractors" or incorrect pictures available for choice. Children also had to deal with the inflection-^ verb skill as they selected their answers. Task 2 was done better by the self -instruction group. The third task involved recognizing vocabulary words and inserting them into sentences and crossword puzzles. The self-instruction group did better on this task. On Task 4, reasoning skills were again required. Children had to use numbers to arrange pictures in a logical sequence, and, also, had to read a passage and then sequence its events. The self-instruction group did better than the no-selfinstruction group. Task 5 involved decoding verbs with and without the inflected-_s ending and putting those words into sentences

PAGE 68

-60that made sense. The self-instruction group performed better than the no-self-instruction group on this task. Task 6 required that students insert vocabulary words into sentences and that they do a crossword puzzle. On this task, on which the self-instruction group performed marginally better (but not to a statistically significant degree) , there were two factors which may have confounded the children. One, on the vocabularyinsertion task, some answer words could be used more than once, which broke an implied rule — "use each word once." Two, the crossword puzzle required answers in both the "across" and "down" sets; the children had previously had to insert words in one direction or the other. Each of these differences was subtle. This was the only task on which the nesting effect was significant. It is possible that, perceiving the format difficulties of this task, some teachers extended directions help beyond the requirements of the basal. One question that occurred as this study was concluded was whether the self-instruction group outperformed the noself -instruction group by virtue of being better readers: Had something happened which simply increased their reading skill? For this reason, the curriculum resource teachers at each school were again contacted and asked for each child's mastery score on the Ginn reading test for Level 3. This mastery test is teacher-dictated,

PAGE 69

-61so that the children know, item by item, what to do. There is no problem identification or strategy selection required. On the mastery test there is no significant difference (at a < .05) between the self-instruction and no-selfinstruction groups, as evidenced by the t-test in Table 4-4. Table 4-4 Scores on Ginn Level 3 Mastery Test M SD t P Experimental (N = 29) 44. 68 1. 94 -1.6 . 10 Control (N = 27) 43.62 2. 73 Alpha set < .05.

PAGE 70

CHAPTER V SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS Summary The purpose of this study was to apply the cognitivebehavior modification self-instruction model to classroom practice and content. Cognitive modeling and children's private speech were combined to develop children's selfguiding statements and images as they engaged in independent work from their reading curriculum. From a review of the literature on the use of cognitive-behavioral strategies with children, the following research needs were indicated: 1. The use of cognitive-behavior behavior with normal subjects on academic tasks in natural settings must be tested. 2. Some research must focus on the processes rather than the product (s) of learning. 3. The capacity to obtain increased generalization or transfer of learning through cognitive-behavior modification must be tested. 4. There must be greater specificity of the independent variable in cognitive-behavior research. -62-

PAGE 71

-635. The impact of cognitive-behavior modification as a treatment in classroom and home settings has not been tested. 6. There is a need to show that metacognitive changes are reflected in normal performance. Fifty-six first grade students (within nine intact reading groups in five schools) were randomly assigned to two treatment conditions. Each reading group was within a normal first grade classroom in a public school; although some schools had more than one experimental group, no school contained both an experimental and control group, and no classroom had more than one participating reading group. There were three urban and two rural schools, with one of the rural schools in each treatment condition. The treatment condition consisted of self-instruction training using non-reading materials (picture association tasks, mazes, counting tasks) imposed separately from reading instruction. The control group received reading instruction only. The dependent measure for this study was children's performance on the usual independent seatwork tasks provided with the reading series; those tasks were selected and removed from the children's daily workbooks and totaled 85 items in the areas of phonics, word identification in context, and sequencing. The tasks were grouped into six sets for data collection according to the pacing requirements for reading instruction.

PAGE 72

-64A multivariate analysis of variance (MANOVA) on a two-way nested design was used to compare overall group performance across tasks; this analysis showed a very high difference between experimental and control groups (F= 191.99, p = .0052). Six separate univariate ANOVAs were used to determine if the overall difference between groups existed for each task, and these analyses did show a measurably significant difference between groups on five of the six tasks. Conclusions Overall, the results of this study indicate that the cognitive-behavior modification technique of selfinstruct ion is effective and appropriate as a means of improving the academic task performance of normal children. It is important to note that the positive results here were obtained with groups of children in naturalistic settings and on specific academic tasks done without teacher assistance. For the self-instruction group to do significantly better indicates that cognitive strategy changes can be seen on measures of ordinary performance, and children's use of study time could be improved through development of skills in self. monitoring .

PAGE 73

-65Children trained in self-instruction applied academic skills with significantly higher accuracy through the development of planful behavior. Since self-instruction training was implemented with pictorial or reasoning tasks rather than reading items, its successful use in the reading area by students reflects its focus on processes or executive skills and demonstrates its generalizability across some tasks. The results of this study reinforce the appropriateness of the practical application of self-instruction: Teachers were easily and effectively trained to incorporate it into their curricula, and students' performance on academic tasks directly from the district-based curriculum improved. Implications The results of this study have implications for further research on the "pedagogical potential" of cognitive-behavior modification. If self-instruction is instrumental in producing a positive change in the independent work and skills application of young students over a period of a few weeks, its continued (year-to-year) use could possibly establish a repertoire of metacogn i ti ve or executive skills students put to use as they work.

PAGE 74

-66As students improve processes or strategies with which to approach problems, attendant positive changes in products (test scores, problem solving) should be apparent and might be compared to similar successes attained through methods from the area of metacognition. Also, as executive processes improve, their generalizability and durability over time could be examined. Eventually, attention could be given to finding the critical elements within the selfinstruction process to find those which directly apply to school curricula and/or to certain age groups. The effective use of self-instruction in a study such as this one but with children over seven remains to be done; once children have begun to internalize speech, adjustments in the type of selfinstruction training might have to be made. Most importantly, the successful results of this study were due to (a) careful adherence to Meichenbaum ' s guidelines for effective self-instruction; (b) thorough task analysis and expert knowledge of the students engaged in task performance; and (c) thoughtful, patient attention to the development of cognitive processes. Replication or further application of this work should . cautiously include those cares.

PAGE 75

-67For education and for children, successful applied use of self-instruction has other implications. One direct result of self-instruction seems to be students' increased independent behavior by virtue of being able to identify problems better and then plan behavior {or perhaps it is their new knowledge that they are expected to do those things rather than wait for the teacher). Although concern here was for an increase in normal class-task performance, it is logical that such behavior would carry over to tests, homework, and projects assigned to be done independently. It would be ideal for educators to have methods for emphasizing cognitive processes which would, in turn, positively affect the learning products. Educators have spent time and money to raise standards of education through program evaluation, teacher evaluation, examination of pupil/teacher ratios and funding levels; the results that can be obtained with self-instruction suggest that working with students on cognitive strategies may help them make better use of resources currently available. Recommendations Given the success of self-instruction on specific classwork tasks in the reading area by first graders, the obvious expansions of this investigation would be as follows:

PAGE 76

-681. Replicate this study with tasks from other content areas. 2. Compare classes with and without selfinstruction in a year-long study. 3. Follow classes with and without self-instruction into the next grade, looking for generalization and stability of strategy learning. 4. Train children in self-instruction and investigate their use of it across subject matter and situations. 5. Expand the five-to-seven year old age category utilizing self-instruction in academics to see if internalization of speech hampers the use of self-instruction or visa versa. This researcher has particular interest in integrating a full-year, full-class self-instruction regime into the curriculum for a kindergarten group, working long-term to fully automatize the self-instruction strategies, followed by comparison of those students' work in first grade to students without self-instruction training.

PAGE 77

APPENDIX A INVITATION TO PARTICIPATE Dear First Grade Teacher, I am a kindergarten teacher at Stephen Foster and a doctoral student in educational psychology at the University of Florida. I would like to invite you to be part of the research project for my dissertation; the topic is teaching cognitive strategies to young children . You will find this directly relevant to current interest in problemsolving, developing independent work habits, and teaching thinking skills. I am looking for teacher volunteers and am primarily interested in using students in Ginn Level 3. Here is a brief overview of what involvement would mean (using Level 3 as an example) : Week 1 ; Begin Level 3 instruction as usual. Week 2 ; Continue Level 3 instruction as usual. Week 3 ; All continue Level 3 instruction. Those in the experimental group will provide four fifteen-minute cognitive strategy lessons for their students. Those in control groups will teach as usual. Week 4 : Continue Level 3 instruction as usual. Experimental group: Distribute the independent seatwork tasks to your students, reminding them that the strategies they have learned might be good to use. Control group: Distribute the independent seatwork tasks to your students. Week 5: Same as previous week.

PAGE 78

-70All work for both groups will be provided. That work comes directly from the Ginn workbooks and will be organized, stapled, and delivered, along with a pacing schedule keyed to the Ginn manual. You would not need to check any experimental or control work. Errors are expected from both groups. No individual or class data will be used or reported; interest is in comparing the performance of the experimental and control groups which will be comprised of students from several schools. TRAINING Before the third week of instruction, teachers randomly assigned to the experimental group will be trained in teaching cognitive strategies and provided with all materials. Those in the control group will meet (separately from experimental) and receive procedural training and prepared materials. Training will be fun (!), very informative, and brief (maxiiri*um two hours for experimental, a little less for control). At the end of data collection, training will be provided for control group teachers and results will be shared with everyone. The strategies you learn to teach your students will work with all subject areas, are pleasing to the children, and enhance the independent work level in your class. Please consider joining this study! Thanks, If you have Level 3 students in your first grade, and are willing to participate in this study, please notify your curriculum resource teacher or call me at Stephen Foster (372-4363) or at home (378-3992).

PAGE 79

APPENDIX B SELF-INSTRUCTION TRAINING In self-instruction training, cognitive modeling and private speech are combined as an approach to teaching thinking skills. Children are taught how to use selfstatements and images to think and plan behavior; they are not instructed in what to think or encouraged to focus on "right answers." Although right answers would appear to be a logically correct educational goal, both metacognitive and self-instruction researchers reject that aim and set the development of underlying thinking skills or executive processes as a priority of teaching efforts. Self-instruction in task performance requires that the student engage in problem definition, appropriate strategy selection, self-evaluation of performance, selfreinforcement, and coping behaviors. To date, the selfinstruction process has been used with young children on social, interactive, motor, and academic tasks. Meichenbaum's model includes the following steps: 1. cognitive modeling (an adult performs the task and talks to self) ; 2. child performs the task with model's direction for guidance (overt external guidance) ; -71-

PAGE 80

-723. child performs task and talks aloud to self (overt self-guidance) ; 4. child whispers and performs task (faded, overt self-guidance) ; 5. child performs task via inaudible or private speech or nonverbal self-instruction (covert self-instruction). (Meichenbaum, 1977, p. 32) At step three the child's skills in problem definition, planful behavior, and implementation of strategies can be observed. Meichenbaum begs the question of self-instruction research in academics by stating, We can conceive of academic tasks where the teacher provided the children with a set of tasks . . . and the child's job was to identify what the problem is, how he or she will go about solving the task, where the likely pitfalls are, etc. . . . Teachers could give assignments and ask the children to describe in detail how they are going to go about performing the assignment. . . . Discussion could center on the process, not only the product, of the assignment. (Meichenbaum, 1985, p. 421) The purpose of this study is to apply the cognitivebehavior modification self-instruction model as developed by Meichenbaum to actual classroom practice and content. Teachers participating in the experimental group of this study will supplement regular reading instruction with students' training in self-instruction. The student training will involve four sessions of approximately 15 minutes during which the teacher will, through cognitive modeling and the use of the children's own speech.

PAGE 81

-73develop self-guiding statements based on Meichenbatun' s model. Using that overtto-covert style, the self-guiding statements for each group will follow the pattern of 1. problem definition and strategy selection ("Hmmm. What do I do here?"); 2. strategy implementation ("Am I doing what I said I would do?"); 3. error management ("What if I make a mistake, how can I fix it?" ) ; and 4. self-reinforcement/self-checking ("How did I do?" or "Let me check my paper before I give it to the teacher."). Measures of normal performance will be taken on a set of reading tasks from the Ginn studybook and skillpack. Our interest is in the children's use of selfinstruction, not in individual scores or , teacher performance. Errors are, as usual, expected, and the teacher should feel comfortable in accepting the children's independently done work without correcting it.

PAGE 82

APPENDIX C GUIDELINES FOR EFFECTIVE SELFINSTRUCTION Ideally, the teacher should model a metacognitive perspective, one which indicates openness to thinking and anticipation of mistakes which, when they occur, are viewed as problems to be solved. The instructor should be careful of attributions made when failures occur, with statements directed at nurturing the problemsolving attitude. To maximize generalization, training in self-instruction must be on a task which has the same requisites or elements of the performance task, but is overtly different from that performance task. The pupil must be a collaborator in generating the strategies to be used, and the private speech trained must be developed with and compatible with the child's natural style. The cognitive strategy training should take place after component skills and experience with the task are in the child's repertoire, and the pupil must be helped to recognize the new task as one facilitated by self-instruction and requiring the transfer of those strategies. The instructor needs to ensure against rote repetition of patterned strategies, encouraging personal involvement by the student by use of -74-

PAGE 83

-75-1 faded cues and acceptance of the child's wording. The teacher should directly encourage the child to generalize strategies to various tasks or situations, perhaps by engaging in discussion of other suitable uses. Lastly, a sense of self-satisfaction should be nurtured in the child, with attention given to developing the child's self-reinforcing and coping skills as flexibility in the use of strategies is fostered.

PAGE 84

APPENDIX D MEICHENBAUM' S SELFINSTRUCTION MODEL Meichenbaum' s self-instruction model includes the llowing steps: 1. cognitive modeling (an adult performs the task and talks to self) ; 2. child performs the task with model's direction for guidance (overt external guidance) ; 3. child performs task and talks aloud to self (overt self-guidance) ; 4. child whispers and performs task (faded, overt, self-guidance); 5. child performs task via inaudible or private speech or nonverbal self-instruction (covert self-instruction). (Meichenbaum, 1977, p. 32) -76-

PAGE 85

APPENDIX E INSTRUCTIONS FOR THE EXPERIMENTAL GROUP Experimental Group Look over the chart for Level 3 indicating the pages that will be used in the study (attached) . Please discard those pages from your student books. During the third week of instruction, plan four 15-minute Training Sessions in self-instruction. These should be separate from reading group. Do not make any reference to reading. Training Sessions The training sessions are used to elicit the four self-instructional questions from your particular group. Use your reminder card (pictured below) for guidance in structuring the questions, but accept your own children's wordings. Work to get four specific questions, in general langauge (not a rote phrase) that appeals to your group. Introduce this in the following way: "When you are in school you have a lot of jobs to do. Sometimes you need help with directions. Other times you can figure out what to do by yourself. Let's start a Thinking Class together, and look at some work to see if, together, we can ask some questions that will help you to figure out what to do on your own." Use Training Set 1 What is the problem on the page? How can you mark the answer? Sometimes when you are working, the teacher comes around and checks to see if you are working correctly. Is that something you could do for yourself?

PAGE 86

-78How? What if you make a mistake? What would people who are thinking want to do? What usually happens when you turn your work in to the teacher? What does she find? Is that something you could do for yourself? Teacher guidance and summation: Let's go back and see what happens in Thinking Class. First you found the problem on the page — is there a question you can think of to help find the problem? "Hmmm. What do I do here?"* After you figure out what to do, how can you be your own teacher and see how your work is going? "Am I doing it?"* What if you make a mistake, what can you do to fix it and make it OK with yourself? "Am I messing up? What can I do?"* What is a question you could ask yourself in order to be your own "checker" so the teacher won't find so many mistakes? "How did I do?" "Let me check this."* *Samples of answers developed by other students in selfinstruction. Your students' self-guiding statements or questions will reflect their mutual language and style. Repeat the process, streamlining it, over the next three training sessions.

PAGE 87

-79Data Collection After the last training session, in a reading group time, give children Set 1 of the performance tasks. To receive this set, they should be through the story "Look at Fish. " Spend a minute remininding them of selfinstruction questions before the set is distributed. Have them do Set 1 independently; take up the work and go on with reading group as you wish. Do not help the children with their independent work or worry about their mistakes. Continue instruction according to the Ginn manual. There are five data collection points left. Fit them in in the following order: After "What Grandma Does" — Set 3 After "Mix and Make" — Set 5 After "Bread to Eat" — Sets 4 and 6_ (in two separate sessions) After "The Hen and the Bread" — Set 2 Put each child's work in a pile of six sets and call me upl I will take it away and see what happened and will be back in touch with you. Thanks, Additional Administrative Procedures I will place a call to you weekly to check your pacing and progress. I will ask for your training session times and your reading group times for the upcoming week, and answer any questions you have. Additionally, observation times during the term of the study will be planned, and I may drop in at the training session or reading group times.

PAGE 88

-80Reminder Card (You will have this card, laminated, to keep with you to channel your students' questions.) PROBLEM DEFINITION AND STRATEGY SELECTION "Hmmm. What do I do here?" IMPLEMENTATION "Am I doing what I said I would do?" "Am I doing it?" ERROR MANAGEMENT "What if I make a mistake, how can I fix it?" "Am I messing up?" "Ooops! I can cry or I can fix it!" SELF-REINFORCEMENT/SELF-CHECKING "How did I do?" "Let me check my paper before I give it to her! "

PAGE 89

-81Level 3 Pacing Sheet Here is a list of the stories from Level 3. The skillpack (SP) and studybook (SB) pages I am interested in are underlined. Text Story U -7 Ken and the Fish SP 1 2 3 4 5 6 SB 4 5 6 7 8 1 -3 J. ^ Ana and the Whale SP 8 9 10 Tl 12 SB 9 10 11 12 13 14 15 1 4What Is It? SP 13 14 15 16 17 18 SB 16 T7 18 19 20 21 22 ^ u Here Come the Whales 2226 Look at Fish SP 19 20 21 22 23 24 25 SB 23 24 25 26 27 28 29 3035 What Grandma Does SP 25 26 27 28 29 30 31 SB 30 31 32 33 34 35 36 3643 Mix and Make SP 32 33 34 35 36 37 28 SB 37 38 39 40 41 42 4448 Clay to Bake SP 39 40 41 42 43 44 45 SB 43 44 45 46 47 48 49 4955 Bread to Eat SP 46 47 48 49 50 51 52 SB 50 51 52 53 54 55 56 5663 The Hen and the SP 53 54 55 56 57 58 59 Bread SB 57 58 59 60 61 62 63

PAGE 90

APPENDIX F INSTRUCTIONS FOR THE CONTROL GROUP Control Group Look over the chart for Level 3 indicating the pages that will be used for the study (attached) . Please discard those pages from your student books. Data Collection After finishing the story "Look at Fish" distribute task Set 1 toyour students to be done independently. When they finish, take up the work and go on with reading group as you wish. Do not help the children with their independent work or worry about their mistakes. Continue instruction according to the Ginn manual. There are five data collection points left. Fit them in in the following order: After "What Grandma Does" — Set 3 After "Mix and Make" — Set 5 After "Bread to Eat" — Sets 4 and 6^ (in two separate sessions) After "The Hen and the Bread" — Set 2 Put each child's work in a pile of six sets and call me up! I will take it away and see what happened and will be back in touch with you. Thanks, Additional Administrative Procedures I will place a call to you weekly to check your pacing and progress. I will ask for your reading group times for the upcoming week and answer any question you have. Additionally, observation times during the term of the study will be planned; I may drop in at the reading group times. -82-

PAGE 91

-83Level 3 Pacing Sheet Here is a list of the stories from Level 3. The skillpack (SP) and studybook (SB) pages I am interested in are underlined. Text Story 6-9 Ken and the Fish SP 1 2 3 4 5 6 SB 3 4 5 6 7 8 10-13 Ana and the Whales SP 7 8 9 10 11 12 SB 9 10 11 12 13 14 15 14-19 What Is It? SP 13 14 15 16 17 18 SB 16 17 18 18 20 21 22 20-21 Here Come the Whales 22-26 Look at Fish SP 19 20 21 22 23 24 25 SB 23 24 25 26 27 28 29 3035 What Grandma Does SP 25 26 27 28 29 30 31 SB 30 31 32 33 34 35 36 3643 Mix and Make SP 32 33 34 35 36 37 38 SB 38 38 39 40 41 42 4448 Clay to Bake SP 39 40 41 42 43 44 45 SB 43 44 45 46 47 48 49 4955 Bread to Eat SP 46 47 48 49 50 51 52 SB 50 51 52 53 54 55 56 5663 The Hen and the SP 53 54 55 56 57 58 59 Bread SB 57 58 59 60 61 62 63

PAGE 92

APPENDIX G RESEARCH DESIGN ^1 ^2 ^1 ^2 ^3 ^4 ^5 ^6 1 ^6 ^1 ^2 ^3 ^4 ^5 ^6 '^5 ?36 1 4/ ^40 ^2 ^7 1 1 4^16 ^6 f41 1 4^43 '^3 fl7 1 ^24 T-7 S. . 7 , 44 1 4' ^48 '^4 p25 1 4* ^35 ^8 p4 9 1 4/ ^57 ^9 p58 1 4^ ^63 X, are the six task scores for each child Sg^ are the children (subjects) T-j^ Tg are the teachers ^1~'^2 experimental and control conditions -84-

PAGE 93

REFERENCES Bandura, A. (1969). Principles of behavior modification . New York: Holt, Rinehart and Winston, Brown, A. (1980) . Metacogn iti ve development and reading. In R.J. Spiro, B.C. Bruce, & W.F. Brewer (Eds.), Theoretical issues in reading comprehension: Perspectives from cognitive psychology, linguistics, artificial intelligence and education . Hillsdale, New Jersey: Lawrence Erlbaum Associates. Clymer, T, , Venezky, R.L., Johnson, D.D., & Pearson, P.D. (1984). Fish and not fish. Lexington, Massachusetts: Ginn and Company. Craighead, W.E., Craighead-Wilcoxon , L. , & Meyers, A. (1978). New directions in behavior-modification with children. In M. Hersen, R.M. Eisler, & P.M. Miller (Eds.), Progress in behavior-modification. Vol. 6 (pp. 159-201) . New York: Academic Press. Craighead, W.E. (1982) . A brief clinical history of cognitive-behavior therapy with children. School Psychology Review , 11.(1), 5-13. Craighead, W.E., Meyers, A.W., Craighead, L.W., & McHale, S.M. (1983). Issues in cognitive behavior therapy with children. In M. Rosenbaum, C.M. Franks, & Y. Jaffe (Eds Perspectives on cognitive therapy in the eighties (pp. 2 261). New York: Springer. Gagne, R. (1964). Problem solving. In A. Melton (Ed.), Categories of human learnings (pp. 293-317) . New York: Academic Press. Grimm, J., Bijou, S., & Parsons, J. (1973). A problemsolving model for teaching remedial arithmetic to handicapped young children. Journal of Abnormal Child Psychology , 26-39. -85-

PAGE 94

-86Hobbs, S., Moguin, L., Tyroler, M. , & Lahey, B. (1980). Cognitive-behavior therapy with children: Has clinical utility been demonstrated? Psychological Bulletin , 87 , 147-165. Kazdin, A. (1982). Current developments and research issues in cognitive-behavioral interventions: A commentary. School Psychology Review , 1_1 (1) , 76-82. Kendall, P. (1977). On the efficacious use of verbal self-instruction procedures with children. Cognitive Therapy and Research , 1, 331-341. Lovitt, T. , & Curtis, K, (1968). Effects of manipulating antecedent event on mathematics response rate. Journal of Applied Behavior Analysis , 1_, 329-333. Meichenbaum, D. (1977) . Cognitive-behavior modification : An integrative approach . New York: Plenum Press, Meichenbaum, D. (1985) . Teaching thinking: A cognitivebehavioral perspective. In S. Chipman, J. Segal, & R. Glaser (Eds.), Thinking and learning skills (Vol. 2): Research and open questions (pp. 407-426) . Hillsdale, New Jersey: Lawrence Erlbaum Associates. Meichenbaum, D., & Asarnow, J. (1979). Cognitive-behavior modification and metacognitive development: Implications for the classroom. In P. Kendall & S. Hollon (Eds.), Cognitive-behavior interventions: Theory, research and procedures (pp. 11-35). New York: Academic Press. Meichenbaum, D. , & Burland, S. (1979). Cognitive-behavior modification with children. School Psychology Digest , 8.(4), 426-433. Meichenbaum, D. , & Burland, S. (1981). Cognitive-behavior modification with children. In H.F. Clarizio, R.C. Craig, & W.A. Mehrens (Eds.), Contemporary issues in educational psychology (4th Edition) (pp. 109-114). Boston: Allyn and Bacon. Meichenbaum, D. , & Goodman, J. (1971). Training impulsive children to talk to themselves: A means of developing self-control. Journal of Abnormal Psychology , 77 , 115-12 Meyers, A.W., & Craighead, W.E. (1984). Cognitive behavior therapy for children . New York: Plenum Press.

PAGE 95

-87Pressley, M., & Levin, J.R. (1983). Cognitive strategy research: Educational applications . New York: Springer-Verlag . Rhodes, D, (1979). Cognitive self-instruction to increase comprehension in early readers (Doctoral dissertation. University of Louisville, 1979) . Dissertation Abstracts International , 41/02A, 612. Sullivan, F.J. (1981). A comparison between attack strategy training and attack strategy training in combination with self-instruction in teaching academic tasks to first graders (Doctoral dissertation. University of Oregon, 1981). Dissertation Abstracts International , 42/08A, 3520. Vygotsky, L.S. (1962) . Thought and language . Cambridge, Massachusetts: Massachusetts Institute of Technology Press.

PAGE 96

BIOGRAPHICAL SKETCH Christopher Williams Sparks was born on March 23, 1946, in Shreveport, Louisiana, second in a family of six children. She grew up in university towns and settled in Gainesville, Florida, as a result of her family's migration there. Chris has a Bachelor of Arts in Education and a Master of Education from the University of Florida, with specialization in early childhood. She will receive a Doctor of Philosophy in foundations of education in May, 1986, from the University of Florida. Chris has been a teacher in the Alachua County, Gainesville, Florida, school district for 18 years. Although she has a preference for kindergarten-age people, she has also enjoyed her part-time employment in the Childhood Education department at Gainesville's Santa Fe Community College. Post-graduation plans for Chris include further research in cognitive-behavior modification, publication of several articles which have been on the "back burner," and continued teaching. Additionally, she intends to reacquaint her patient children Judd and Sara, who have -88-

PAGE 97

-89been eating a lot of frozen pizza, with good food and motherhood.

PAGE 98

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Foundations of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. John M. Newell Professor of Foundations of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Linda M. Crocker Professor of Foundations of Education

PAGE 99

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. This dissertation was submitted to the Graduate Faculty of the College of Education and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May 1986 Associate Professor of instruction and Curriculum Chairman, Foundations of Education Dean, College of Educ Dean, Graduate School