Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A meta-analysis of Feuerstein’s Instrumental Enrichment Shiell, Janet Lillian 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2002-732460.pdf [ 8.27MB ]
Metadata
JSON: 831-1.0055041.json
JSON-LD: 831-1.0055041-ld.json
RDF/XML (Pretty): 831-1.0055041-rdf.xml
RDF/JSON: 831-1.0055041-rdf.json
Turtle: 831-1.0055041-turtle.txt
N-Triples: 831-1.0055041-rdf-ntriples.txt
Original Record: 831-1.0055041-source.json
Full Text
831-1.0055041-fulltext.txt
Citation
831-1.0055041.ris

Full Text

A M E T A - A N A L Y S I S OF FEUERSTEIN'S I N S T R U M E N T A L E N R I C H M E N T BY JANET LILLIAN SHIELL B . E d . (Secondary), The University o f British Columbia, 1976 M . E d . (Counselling Psychology), The University o f British Columbia, 1984 A THESIS SUBMITTED IN P A R T I A L F U L F I L M E N T OF T H E REQUIREMEN15FOR T H E D E G R E E O F D O C T O R OF PHILOSOPHY In T H E F A C U L T Y OF G R A D U A T E STUDIES Centre for the Study o f Curriculum and Instruction W e accept this thesis as conforming to the required standard  T H E U N I V E R S I T Y OF BRITISH C O L U M B I A A p r i l 2002 © Janet L i l l i a n Shiell, 2002  In  presenting this  degree at the  thesis  in  University of  partial  fulfilment  of  of  department  this thesis for or  by  his  or  requirements  British Columbia, I agree that the  freely available for reference and study. I further copying  the  representatives.  an advanced  Library shall make it  agree that permission for extensive  scholarly purposes may be her  for  It  is  granted  by the  understood  that  head of copying  my or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department of The University of Vancouver, Canada  DE-6 (2/88)  Abstract The purpose o f this study was to examine the effectiveness o f Instrumental Enrichment, an educational program developed by Reuven Feuerstein, which attempts to improve an individual's ability to reason. A meta-analysis was performed on studies that had been conducted between the years 1979 to 1996. Thirty-six studies were analyzed according to their results from measurements in the cognitive/visual-perceptual, academic achievement and affective domains. The results from the meta-analysis were mixed. There were significant combined effect sizes o f 0.24 for non-verbal ability, 1.41 for verbal ability, and 0.60 for one combination o f full-scale ability. Significant effect sizes for measures o f visual perception and visual-motor ability were 0.42, 0.71 and 1.68. There were also significant effect sizes for general achievement and for one combination o f math achievement at 0.26 and 0.29 respectively. The intellectual achievement locus o f control effect size was significant at 0.33. (Cohen (1988) considers an effect size o f 0.20 to be small, that o f 0.50 to be o f medium size, and that o f 0.80 to be large.) A l l effect sizes for reading were nonsignificant as was that for the Learning Potential Assessment Device. A l l other measures, such as those for self-confidence, self-concept, motivation and attitudes, in the affective domain were nonsignificant. One effect size for motivation and attitudes was significantly negative, indicating that the control group outperformed the treatment group. The results are discussed in terms o f near- and far-transfer o f learning.  ii  Table o f Contents Abstract Table o f Contents List o f Tables Acknowledgements Chapter One: Cognitive Orientations to Curriculum Theoretical Background The Beginnings: Piaget's Stages and Vygotsky's Zone o f Proximal Development Theories o f Critical Thinking Instrumental Enrichment Structural Cognitive Modifiability Mediated Learning Experience ( M L E ) Chapter T w o : Literature Review o f Feuerstein's Instrumental Enrichment Goals Curriculum Literature Review Individual Studies Groups o f Studies Intelligence Summary o f Evaluations Importance o f the Study Chapter Three: Literature Review o f Meta-Analysis Definitions Usefulness • Critique Theory Development Type II Errors Validity Reliability Homogeneity Sources o f Variation and Generalizability Bias The File-Drawer Problem Quality Multiple Publications Sample Size Studies W i t h More Than One Statistic Unbiased Estimate o f Effect Size Procedure Literature Search Inclusion-exclusion criteria Data Collection Statistical Analysis Choosing Estimates o f Effect iii  1 1 1 1 2 2  Calculations for the Unbiased Estimator o f Effect Size for Each Study Interpretation o f Effect Sizes  55 56  Chapter Four: Methodology Methodology Criteria Used to Include or Exclude Specific Studies Quality o f Individual Studies Statistical Methodology The File-Drawer Problem The Hawthorne Effect  •  Chapter Five: Results Overall Effect Sizes Near and Far Transfer  58 58 58 61 63 64 64 •• 66 66 67  Chapter Six: Discussion Near Transfer Mid-Transfer Far Transfer Subgoals o f Instrumental Enrichment  71 73 74 75 .76  Chapter Seven: Implications for Practice and Research Implications for Practice Affective Domain Instruction Counselling Academic Instruction and IE Implications for Research The L P A D Teacher Effectiveness and Bridging Further Analysis o f Data in this Meta-Analysis Brain Research Limitations o f this Study  78 78 78 79 80 81 81 82 83 83 84  ~.  References  86  Appendix Appendix Appendix Appendix Appendix Appendix  96 142 146 152 156 161  A : Summary o f Studies Used, Measures and Effect Sizes B : Summary o f Studies Not Used C : Results from Random Effects Analysis D : Significant Results Information E : Nonsignificant Results Information F : Effect Sizes from F-Tests and Reported Effect Sizes  iv  ,  List o f Tables 1. Summary o f Included and Excluded Studies 2. Summary o f Transfer Results  61 70  v  Acknowledgements I would like to express my deep appreciation to the members o f my doctoral committee for their guidance and support. Thanks to Dr. H i l l e l Goelman for shepherding the dissertation process through university administration, and for his encouragement and humour when I despaired. Thanks to Dr. Michael Schulzer for his patience while explaining many statistical methods and formulae. The meta-analysis was a complicated procedure and Michael's assistance was invaluable. Thanks also to Dr. M a r i l y n Samuels for her insightful comments regarding the Instrumental Enrichment program and research related to it.  yi  Chapter One Cognitive Orientations To Curriculum The purpose o f this study is to examine the effectiveness o f an educational program, developed by Reuven Feuerstein and colleagues, which attempts to improve an individual's ability to reason. Trained as a clinical psychologist, Feuerstein became involved i n the psychological assessment o f Jewish children who were being resettled in Israel after W o r l d War II. M a n y had experienced poverty and deprivation, and many were Holocaust survivors. The majority o f these children also scored on standardized intelligence tests i n the retarded range. Because o f these circumstances, Feuerstein began to reconsider traditional notions o f intelligence and its measurement. He also began to explore the development o f an educational intervention to enhance cognitive functioning during adolescence. That intervention, the focus of this study, is known as "Instrumental Enrichment" (Feuerstein, K l e i n , and Tannenbaum, 1991; Feuerstein, Rand, M . B . Hoffman, & Miller, 1980; Yates, 1987). Although there are many programs designed to encourage or improve cognitive development, I have chosen Feuerstein's Instrumental Enrichment (FIE or IE) for several reasons. In the early 1990's, I was trained to teach the first half o f the Instrumental Enrichment program and to administer its related assessment tool, the Learning Potential Assessment Device. I was very impressed by the methods used and the enthusiastic acceptance o f them by participants in the learning groups. I reviewed the literature to see what had been written about the effectiveness o f F I E and found a mixture o f results. Most authors, at least on an intuitive level, believed the program was, or could be, effective, but there seemed to be more success in some areas than in others. I did a small meta-analysis in 1993 to explore FIE's overall effectiveness and found that it appeared to be most successful when measured with assessment instruments closest to the material used in the program, and it did not seem to be  1  effective at all i n the affective domain. There were methodological difficulties in the execution o f my analysis, however, and I wanted to do a more thorough and carefully designed meta-analysis with as many studies, published and unpublished, as I could obtain. This w i l l be an exploratory, descriptive study, not meant to prove or disprove the effectiveness o f F I E or the theory behind it. I had no a priori hypotheses at the beginning o f the study; the research questions arose from the literature review. Before discussing Feuerstein's theory and methods in detail later i n this chapter and i n Chapter T w o , I w i l l first describe the contributions o f other theorists toward our understanding of intelligence and the process o f learning. This is so we can be aware o f the background within which Feuerstein formed his ideas. The first half o f the chapter w i l l present theories under the sub-headings: The Beginnings and Theories o f Critical Thinking. The second half o f the chapter w i l l present an introduction to F I E and a presentation o f Feuerstein's theories o f Structural Cognitive Modifiability and the Mediated Learning Experience ( M L E ) .  Theoretical Background The Beginnings: Piaget's Stages and Vygotsky's Zone of Proximal Development Jean Piaget. Jean Piaget, whose model encompasses mainly the first 15 years o f life, wrote about cognition i n relation to the developmental process (Piaget, 1952, 1971,1985). H e argued that the major function o f intelligence is to help children adapt to the environment. Children construct their own logical structures through interaction with the environment and move from one stage to another through the process o f equilibration. They experience disequilibrium when they confront aspects o f the environment that cannot be resolved using their currently available schemes. The process o f equilibration, that is, finding a balance between the assimilation o f new information to cognitive schemes (organized patterns o f thinking) and  2  accommodation by the learner to that information, was a means o f resolving cognitive conflict. Cognitive conflict promoted the growth o f personal logic systems by confrontation between new and stored knowledge (Samuels, K l e i n , and Haywood, 1994). Feuerstein was influenced by Piaget's position that the organism, or child, influenced more complex learning than that explained in the Skinnerian, behaviourist view o f direct stimulus-response mechanisms. The Skinnerian model could be illustrated as S->R while Piaget's model was S - » 0 - » R . Feuerstein modified Piaget's model to include a human mediator in his theory o f the Mediated Learning Experience ( M L E ) (Kaniel and Feuerstein, 1989). The theory w i l l be discussed in more detail later in this chapter.  Lev Vygotsky. L e v Vygotsky was a Russian psychologist who believed that all psychological processes are originally encountered within a social context. Children first experience cognitive activities such as problem-solving in the presence o f significant others. They gradually begin to share i n the cognitive activity and to internalize such functions until, eventually, cognitive work can be performed independently (Haywood, B r o w n , and Wingenfeld, 1990; Samuels, et al., 1994). Within Vygotsky's theory o f cognitive development is the concept o f a "zone o f proximal development," originally explained i n 1934, which is the difference between the actual and the potential level o f a child's cognitive functioning (Vygotsky, 1962). Vygotsky's definition was "the distance between the actual developmental level as determined by independent problem-solving and the level o f potential development as determined through problem-solving under adult guidance or i n collaboration with more capable peers" (Vygotsky, 1978, p.86). The notion is that adults can help a child perform tasks and solve problems meaningfully by asking questions, directing attention, providing cues, monitoring 3  performance, and in general, serving as an executive function for the duration o f the task. Over time, children come to perform the tasks by themselves. In this process they internalize the roles taken by their adult helpers and develop the higher order functions that were performed by the adults (Das and Conway, 1992; Secada, 1991). Vygotsky saw the zone o f proximal development as a way o f understanding the kinds o f psychological processes that children would be capable o f in the next phase o f their development and used the analysis o f the zone to identify the type o f instruction needed to realize this potential (Samuels et al., 1994). Although not directly influenced by Vygotsky, Feuerstein was inspired by a colleage o f Piaget's, Andre (Feuerstein and Feuerstein, 1991). In a manner similar to Vygotsky, Rey (1934, 1962) encouraged test administrators to interact with subjects i n order to gain more information about their processes o f learning. This w i l l be discussed further i n the Instrumental Enrichment section o f this chapter. Theories of Critical  Thinking  This section is included because Feuerstein's Instrumental Enrichment is taught according to one o f the critical thinking approaches, and, in Chapters Six and Seven, results o f this meta-analysis are related to critical thinking approaches. The beginnings o f contemporary views o f critical thinking are evident in the work o f a number o f philosophers o f education. John Dewey used the term "reflective thinking" to refer to "the kind o f thinking that consists in turning a subject over in the mind and giving it serious consecutive consideration (Dewey, 1933, p. 3). He stressed the idea o f a problem-solving focus to learning. H e also emphasized the necessity for education to go beyond the teaching o f the subject matter alone and to address the teaching o f thinking.  4  Robert Ennis became highly influential in terms o f promoting interest in, and debate about, critical thinking when he described his conception o f critical thinking in a paper in 1962. H i s somewhat narrow definition, "the correct assessing o f statements" i n that paper has been replaced by a broader one: "reasonable, reflective thinking that is focused on deciding what to believe or do" (Ennis, 1985a, p. 46). In this conception, critical thinking is composed o f both skills (abilities) and dispositions (attitudes). Current usage o f the term "critical thinking" generally reflects Ennis' broad definition. Another theorist, John McPeck, refers to critical thinking as the appropriate use o f reflective skepticism within the problem area under consideration (McPeck, 1981). M c P e c k believes that the criteria that determine the appropriateness o f thinking arise entirely from within particular disciplines or problem areas (Bailin, 1994; Kennedy Fisher, & Ennis, 1991). Within the field, there is debate about whether critical thinking is subject specific or generalizable across subject areas (Bailin, 1994; Kennedy et al., 1991). M c P e c k (1981, 1990) argues against the existence o f general skills because thinking is always about a subject, so general thinking ability detached from a subject cannot conceptually exist. In his view, critical thinking would be taught specific to a particular subject. He is a proponent o f the "immersion" instructional approach in which students become deeply immersed in the subject, and are encouraged to think critically about it, but in which general critical thinking principles are not made explicit. Other authors, such as Swartz (1987), who favour subject-specific instruction follow an "infusion" approach in which, not only is there thought-provoking instruction o f the subject, but general principles o f critical thinking dispositions and abilities are also made explicit (Ennis, 1992).  5  O n the other hand, there are those that maintain that there are general principles o f critical thinking and that these should be taught separately from the standard subject areas. Feuerstein's method follows this approach (Feuerstein et al., 1980). A fourth view o f critical thinking is that it is a combination o f using a set o f general dispositions and abilities along with specific experience and knowledge within a particular area o f interest. Ennis (1985b) and Sternberg (1987) have pointed out that each approach has its advantages and that a mixed model o f providing a separate critical thinking course along with the insertion o f critical thinking approaches into specific subjects should reinforce all skills taught (Kennedy et al., 1991). Part o f the generalizability debate concerns how, or if, critical thinking skills transfer after they have been taught. Transfer refers to the carry-over and use o f skills and knowledge to domains other than the ones in which they were taught. This can mean transfer across academic disciplines or from the academic to the nonacademic world o f the student. M c P e c k (1992) believes thinking is always about a subject; he agrees that some general skills exist, but that their usefulness is limited. The most useful skills, he claims, tend to be limited to narrow domains. For example, some skills, such as counting, may seem quite narrow, but their domains o f application are quite large. Blatz (1992) accepts that people can be taught to think in general ways, but this does not guarantee that they w i l l have learned to think i n ways that are constrained by different community paradigms and information contexts. Several other authors state that whether or not critical thinking is generalizable depends on what critical thinking is. Although Ennis' definition is generally used, there is, nevertheless, extensive debate on the definition o f critical thinking (See, for example, Johnson, 1992; Norris, 1992; Siegel, 1992).  6  Closely related to the discussions among philosophers over the transfer o f critical thinking are those among educators and psychologists over the transfer o f learning. Salomon and Perkins (1989) and Perkins and Salomon (1987) synthesized findings concerned with transfer and speculate on two mechanisms, the "low road" and the "high road" to transfer. L o w road transfer happens when stimulus conditions in the transfer context are similar enough to those in a prior context o f learning to trigger well-developed semiautomatic responses. This is a relatively reflexive process and occurs most often in near transfer. H i g h road transfer depends on thoughtful abstraction from the context o f learning or application to a new one and a deliberate search for connections. It occurs often in far transfer situations. The two roads can work together in that some connections can occur reflexively while others are sought out (Perkins & Salomon, 1994). Perkins and Salomon (1987) also discuss two broad instructional strategies to foster transfer: hugging and bridging. Hugging is used for reflexive transfer. It recommends approximations to the performance desired. For example, a teacher might give trial exams to students rather than just talking about exam technique. The learning experience, therefore, "hugs" the target performance, maximizing the later likelihood o f automatic low road transfer. Bridging is used for high road transfer. The instruction encourages the formulation o f abstractions, searches for possible connections, thoughtfulness and metacognition. For example, a teacher might ask students to create an exam strategy based on their past experience. The instruction would emphasize deliberate abstract analysis and planning (Perkins & Salomon, 1994).  Instrumental Enrichment Feuerstein's Instrumental Enrichment (FIE) belongs within the model o f critical thinking which holds that there are general principles o f critical thinking that can be taught  7  separately from the standard subject areas. I w i l l begin with a brief description o f the program, the history o f its development, and a description o f its main theoretical components, "Structural Cognitive Modifiability" and "Mediated Learning Experience." FIE's purpose, goals and curriculum w i l l be discussed in Chapter Two. The Instrumental Enrichment program is a program consisting o f 13 to 15 "instruments" made up o f paper-and pencil exercises to be completed in 1-hour lessons three to five times a week for two to three years. Each instrument focuses on a specific "cognitive deficiency" and contains exercises that are meant to provide experience in overcoming it. F I E uses abstract, content-free, organizational, spatial, temporal, and perceptual exercises which are meant to involve a wide range o f mental operations and thought processes. The exercises are not intended to substitute for, but to supplement, the traditional content materials o f the regular classroom. According to Feuerstein et al. (1980), the aim o f the F I E program is to change the overall cognitive structure o f "retarded" performers by transforming their passive and dependent cognitive style into that characteristic o f autonomous and independent thinkers. Feuerstein and his colleagues believe that l o w cognitive performance need not be regarded as a stable characteristic of an individual and that systematic intervention, directed at the correction o f deficient functions, w i l l improve the condition by producing a change in the cognitive structure o f the individual. Feuerstein believes that, except i n the most severe instances o f genetic and organic impairment, the human organism is open to modifiability at all ages and stages o f development. M u c h o f the original evidence in support o f Feuerstein's theory o f modifiability derives from the results o f his work with Youth A l i y a h , an institution dedicated to the task o f the ingathering and integration o f Jewish children into Israel. Initially, during the pre-war period o f the 1930's, Youth A l i y a h was involved i n rescuing children threatened by the N a z i regimes i n Germany and Austria. During and after the war, the rescue operation spread throughout Western  8  and Eastern Europe in an effort to reach orphaned children who had survived the Holocaust. Feuerstein began to work with Youth A l i y a h and between 1950 and 1954 he and his colleagues examined large numbers o f adolescents in transit camps in Morocco and southern France. The young people came from many cultures, some non-industrialized, i n A s i a , Africa and Europe. They had to be received, settled, classified and schooled for citizenship in a new country with a unique and thoroughly modern technological culture. Tests o f many kinds were given as a basis for planning their education after immigration. But existing tests proved inadequate to the task because they reflected what the children had failed to learn, not what they could learn, and their achievement status, not their learning potential (Feuerstein, Feuerstein, & Gross, 1997; Feuerstein etal., 1980). Feuerstein was inspired by Piaget and his associate at the University o f Geneva, the clinical psychologist, Andre Rey (Feuerstein & Feuerstein, 1991). In a manner similar to Vygotsky, Rey (1934, 1962) believed that process-oriented tests would reveal information about the processes o f learning and would help assess changes behind adaptive responses. He also encouraged examiners and subjects to interact in order to help discover any processes responsible for low test performance (Haywood, 1977; Haywood et al. 1990). Influenced by these ideas, Feuerstein began work leading to a radical shift from a static to a dynamic approach i n which testing situations were changed into learning experiences for the children. The Learning Potential Assessment Device ( L P A D ) and his instructional approach, cognitive modification with Instrumental Enrichment, developed from this work.  Structural Cognitive Modifiability The F I E program arises from Feuerstein's theory o f "Structural Cognitive Modifiability" (Feuerstein, M i l l e r , M . B . Hoffman, Rand, Mintzker, & Jensen, 1981; Feuerstein, Rand, M a l k a Hoffman, Mendel Hoffman and Miller, 1979); it describes the capacity o f human beings to change  9  or modify the structure o f their cognitive functioning in order to adapt to the changing demands o f life situations. The theory contends that it is useful to view intelligence as a set o f cognitive functions, or modes o f thought. Feuerstein believes it is possible to teach basic cognitive processes after the time o f their normal acquisition, and an essential feature o f this approach is that it is directed not merely at the remediation o f specific behaviours and skills but at changes o f a "structural" nature that alter the course and direction o f cognitive development. Structural changes refer not to isolated events but to an individual's manner o f interacting with, that is, acting on and responding to, sources o f information. Thus, a structural change, once set in motion, w i l l determine the future course o f an individual's development. Cognitive modifiability depends on the individual's ability to learn to react to external and internal stimuli with strategies, that is, methods and general plans, and to be his or her own teacher and modifier. The individual with cognitive modifiability accumulates information, classifies and organizes it, seeks surrounding laws, raises hypotheses and examines them. H e or she makes use of these processes on a wide variety o f subject matters. Cognitive modifiability is considered possible irrespective o f etiology, age or severity o f condition. Feuerstein has described three main characteristics that define structural modifiability. The first is permanence which refers to the endurance o f the cognitive changes over time, or durability. The second is pervasiveness which is related to a "diffusion" process in which changes in one part affect the whole. Finally, the third is centrality which reflects the self-perpetuating, autonomous and self-regulating nature o f cognitive modifiability (Kaniel and Feuerstein, 1989; Tzuriel, 1991; Tzuriel and Haywood, 1992). Cognitive functions are learned i n the normal course o f development, primarily through an adult-directed process called "mediated learning," the chief function o f which is to assist children in organizing perceptual data, inducing and applying rules, monitoring behaviour, and developing  10  intrinsic motivation to learn. Such mediation is claimed to be essential to adequate cognitive development. When adequate mediated learning does not take place, cognitive development, and consequently effectiveness o f learning and problem solving, w i l l be impaired (Feuerstein et al., 1980). Mediated Learning Experience  (MLE)  Although Feuerstein was inspired by Piaget, he noted what he believed were shortcomings to Piaget's theory. According to Feuerstein, in the Skinnerian, behaviourist view, learning occurred by direct stimulus-response mechanisms such as classic pr operant conditioning, or S —» R. B y crediting the organism (i.e., the child), its level o f maturation and its stage o f development with an important role in the accommodation and elaboration o f stimuli, Piaget interposed the organisim between stimulus and resonse, or S - » O —> R. Feuerstein, however, believed that this model underestimated the role played by an adult, or other human (H). H i s S (H) -> O —» R model holds that for optimum learning to occur, a mediator must deliberately place himself or herself between the stimulus and organism with the intention o f altering both (Feuerstein & Feuerstein, 1991; Feuerstein & M . B . Hoffman, 1982; Kaniel & Feuerstein, 1989; Sewell & Price, 1991). A n expansion o f Feuerstein's above model is S (H) - » O <-> (H) - » R. It is described by Jensen, Feuerstein, Rand, Kaniel, and Tzuriel (1988) as follows: The mediator (H) selects stimuli from the environment (S) transforming them according to purposes and goals before they reach the systems o f the learner (O). The mediator (H) selects responses produced by the learner (O) shaping and transforming them to develop response modalities (R). MLE establishes prerequisites and instills need systems required by higher levels o f modifiability. Following mediation the learner (O) is able to interact effectively with the environment (S-O-R) without mediation and adapt to needs to change. (P- 67) 11  Feuerstein calls the human mediation process the Mediated Learning Experience ( M L E ) . According to him, adequate cognitive development cannot take place without the active participation o f mediators. Deficient cognitive functions are seen to be the result o f inadequate mediated learning experience rather than o f social class, poor education o f parents, or psychopathology in children or parents, variables which are often cited by developmental and educational psychologists and sociologists. Although biological, social and environmental factors are acknowledged as possible "distal," or more removed causes o f intellectual differences among individuals and groups, the "proximal" or more immediate cause o f intellectual performance is assumed to be M L E . A lack o f M L E , Feuerstein argues, drastically limits the meaning and significance o f experienced events because these experiences are grasped in an episodic manner without being related to other events experienced by the individual. Therefore, the effects o f a lack o f M L E may be thought o f as depriving individuals o f the prerequisites o f higher mental processes, despite potentially normal capacities inherent in them (Feuerstein et al., 1980; Feuerstein etal., 1981). Feuerstein believes that, through the process o f mediation, culture-characteristic modes o f thought can be passed from one generation to the next. When he was working i n Geneva and then France with the young people from different cultures prior to their emigration to Israel, he came to believe that the differences in test performance between these children and those from Geneva came from two sources, "cultural differences" and "cultural deprivation." While the culturally different individual displayed the capacity to become modified through direct exposure to stimuli, the culturally deprived individual did not. H i s observations, along with those o f Andre Rey, led him to consider the culturally different as individuals equipped with learning capacities which they acquired while being exposed to, and affected by, their own culture. This enabled the individual to benefit from formal and informal opportunities to learn. The culturally deprived, however, had  12  to "learn to learn" v i a mediation. Children i n this group were defined as culturally deprived because they were not exposed to their own culture and therefore became unable to benefit from direct exposure to sources o f stimuli. Their "modifiability" was more or less restricted (Feuerstein and M . B . Hoffman, 1982; Feuerstein and Feuerstein, 1991; Jensen et al., 1988). According to Kopp-Greenberg (1991), many theorists such as Vygotsky (1929) would agree with Feuerstein that mediation occurs naturally as part o f an adult's need to pass on culture from one generation to the next. They believe that cultural transmission through adult transactions usually involves relating cultural values and norms in such a way that they transcend the immediate needs o f the particular circumstances and, therefore, foster mediated learning. According to Feuerstein, Rand, and M . B . Hoffman (1979), Feuerstein, Rand, M a l k a Hoffman, Mendel Hoffman et al.(1979), Feuerstein et al. (1980), Feuerstein, Rand, Jensen, Kaniel, and Tzuriel (1987), Feuerstein and Feuerstein (1991), and Tzuriel and Haywood (1992), there are 11 to 12 characteristics o f M L E , with only the first three considered necessary and sufficient for an interaction to be classified as a mediated interaction: intentionality and reciprocity, meaning, and transcendence. Several other authors, however, believe that the first five or six criteria are important (see, for example, Haywood, 1987; Missiuma and Samuels, 1988). The first six, discussed i n Feuerstein and Feuerstein (1991), Greenberg and Woodside (1994), L i d z (1991), and Tzuriel and Haywood (1992) are described below. Intentionality and Reciprocity.  Intentionality refers to the mediators' intentional efforts  to produce in children a state o f vigilance in order to help them to register some information. Reciprocity refers to the level o f receptivity o f the children to the intentions o f the mediators. Meaning.  This refers to interactions in which the presented stimuli pose affective,  motivational and value-oriented significance. The mediator attaches importance and  13  enthusiasm to the stimuli verbally and nonverbally and the mediation answers the questions o f "why" and "what for." Transcendence.  Transcendence refers to both the character and the goal o f M L E  interactions. The objective is to transcend the immediate needs and specific situation and reach out for goals that might have nothing to do with the original activities. It is any transcendent connection between the content o f the given domain and the content o f some other domain. Mediation offeelings  of competence. The mediator communicates to the children in  various verbal and nonverbal ways that the children are capable o f functioning independently and successfully. Regulation of behaviour.  The mediator controls the system o f the children's response  prior to overt behaviour i n order to inhibit impulsive behaviour or to accelerate the children's activity, depending on the tasks' characteristics and children's responses. It refers to any means the mediator employs to assist children in controlling their approach to a given activity. Mediation of sharing.  Sharing is the energetic component that permits transmission o f  information; it reflects the need o f individuals to move in the direction o f participating with others and to make the others participate with them. The assumption in Feuerstein's theory is that i f our thinking ability is deficient, the reason is mainly due to a lack o f mediated learning experiences. Further, once cognitive deficiencies are identified, they can be strengthened. Biological, social or environmental factors are not as limiting as traditional educators believe. This is the main purpose o f the F I E program. In the next chapter I w i l l describe the program's purpose, as well as its goals and curriculum. I w i l l also review the literature to see how it has been evaluated to date, and I w i l l use this review to develop the questions this study w i l l attempt to answer.  14  Chapter T w o Literature Review of Feuerstein's Instrumental  Enrichment  According to Feuerstein, M . B . Hoffman, Rand, Jensen, Tzuriel, and D . B . Hoffman (1986), a purpose o f Feuerstein's Instrumental Enrichment (FIE) is to enhance the ability o f children to benefit from their direct exposure to new experiences in both formal and informal situations. F I E is a "substitute for early mediated learning experience. It is a phase-specific method and technique for the alleviation o f the i l l effects resulting from the lack o f mediated learning experience..." (Feuerstein et al., 1980, p. 1). One o f the expected results is that children w i l l become more aware of their cognitive processes and abilities. That is, they w i l l increase their metacognitive awareness; this should give them greater control over their cognitive styles and therefore greater consistency in the use o f their cognitive processes. Goals Although the general goal o f F I E is to increase the cognitive modifiability o f learners, there are six specific subgoals that can be regarded as a set o f strategies for achieving the general goal (Egozi, 1991; Feuerstein et al., 1980). >  First is the correction o f any deficient cognitive functions that, i n the theory, are regarded as the result o f inadequate and insufficient early mediated learning experiences. These deficiencies can be in the "input," phase o f information processing where all necessary information is gathered, the "elaboration" phase where the gathered information is used or the "output" phase where the solution to a problem is expressed. The following descriptions are from Egozi (1991), Feuerstein et al. (1997), Feuerstein, Rand, Haywood, M . B . Hoffman, and Jensen (n.d.) and the F I E Curriculum Training Handbook, Vancouver School District (n.d.). Impaired cognitive functions affecting the input level include:  15  •  Blurred, sweeping perception. Stimuli are perceived either incompletely or with few details, a lack o f clarity, and/or imprecise definition o f their borders.  •  Unplanned, impulsive and unsystematic exploratory behaviour.  •  Lack of, or impaired receptive verbal tools and concepts that affect discrimination.  •  Lack of, or impaired spatial and temporal relationships.  •  Lack of, or impaired conservation o f constancy's such as size, shape, quantity, orientation across variations in one or more dimensions.  •  Lack of, or deficient, need for precision and accuracy in data gathering.  •  Lack of, or impaired capacity for considering two or more sources o f information at once.  Impaired functions affecting the elaborational level include: •  Inadequacy in experiencing the existence o f an actual problem and then defining it.  •  Inability to select relevant as opposed to irrelevant cues.  •  Lack o f spontaneous comparative behaviour or the limitation o f its application by a restricted need system.  •  Narrowness o f the mental field. Remembering and keeping i n mind the various pieces of information needed.  •  Lack of, or impaired need for summative behaviour. Summative behaviour makes use o f both absolute and relative quantification in grouping, comparing, subtracting and perhaps multiplying events.  •  Episodic grasp o f reality, that is, viewing each object and/or event as a unique, isolated, one-time phenomenon without any relationship to what has preceded it or may follow.  •  Lack of, or impaired need for pursuing logical evidence.  16  •  Lack of, or impaired interiorization, that is, having a good picture i n one's mind o f what is being looked for or what must be done.  •  Lack of, or impaired inferential hypothetical thinking and lack of, or impaired strategies for hypothesis testing.  •  L a c k of, or impaired ability to define the framework necessary for problem-solving behaviour.  •  Lack of, or impaired planning behaviour.  •  Nonelaboration o f certain cognitive categories because the verbal concepts are not a part o f the individual's verbal inventory on a receptive level, or they are not mobilized at the expressive level.  Impaired functions at the output level include: •  Egocentric communication modalities. This deficiency is a function o f a lack o f differentiation i n which speakers do not perceive the listeners as being different from themselves. The speaker's communication lacks the detail, precision, explicitness and argumentation necessary for the listener to understand the information that is conveyed.  •  Difficulties in projecting virtual relationships, that is, looking for new examples o f relationships that have been learned (bridging).  •  Blocking. This may range from a lack o f initiation o f new responses to open avoidance o f stimuli that may cause the person to react.  •  Trial and error behaviour.  •  Lack of, or impaired verbal tools for communicating adequately.  •  Deficiency o f visual transport, that is, carrying an exact picture o f an object in the mind's eye to another place for comparison without losing or changing some details. 17  >  •  Lack of, or impaired need for precision and accuracy i n communicating one's response.  •  Impulsive, acting out, behaviour affecting the nature o f the communication process.  The second subgoal o f F I E is the enrichment o f cognitive repertoire, including the vocabulary, cognitive operations, strategies, concepts and relationships necessary to complete F I E tasks.  >  The third subgoal is the production o f intrinsic motivation, or the need to learn, through the formation o f new habitual strategies, operations and principles.  >  Subgoal four is the production o f reflective, insightful thinking processes.  >  The fifth is to increase task-intrinsic motivation and to motivate students toward task-oriented abstract goals, that is, the readiness to perceive the performance o f an activity as a goal in itself, solely for the pleasure o f doing it.  >  The sixth and final subgoal is to change students' self-perception from that o f being a passive recipient o f information to that o f being originators o f new information and to encourage their creativity. Curriculum. A complex curriculum has been developed to help teachers use M L E with students. It  consists o f 13 to 15 instruments that are generally taught three to five times weekly for 2 to 3 years or approximately 200 to 350 hours in total. (The hourly amount varies depending on the author). Materials are organized into a series o f units, each o f which focuses on a cognitive function and its related cognitive deficits. B e l o w is a list o f 13 instruments, from L i n k (1991, pp. 9-10), which would cover 3 years o f instruction. Cognitive functions that are addressed for the four instruments in the first year are also given (Vancouver School District, n.d.). In the first-year curriculum, students use the following instruments: 1. Organization of Dots. These materials are intended to help students find the relationships, (i.e., shapes, figures, and other attributes) among a field o f dots, a task similar to  18  picking out constellations in the night sky. In this way students begin developing strategies for linking perceived events into a system yielding comprehensible information that can be a basis for understanding and logical response. The instrument is designed to develop the following cognitive functions: formation o f relationships, representational behaviour, use o f concepts, stable systems o f reference, precise and accurate communication, definition o f the problem, visual transport, use o f two or more sources o f information simultaneously, systematic work, hypothetical/inferential thinking and data gathering. 2. Orientation in Space I. The aim o f this instrument is to promote the creation o f specific strategies for differentiating frames o f reference i n space, such as left, right, front, and back. The materials are designed to develop the following cognitive functions: formation o f relationships, representational behavior, use o f concepts, stable systems o f reference, precise and accurate communication, definition o f the problem, visual transport, use o f two or more sources o f information simultaneously, systematic work, hypothetical/inferential thinking, and data gathering. 3. Comparisons.  This instrument is intended to foster precise perception, the ability to  discriminate by attribute (i.e., equal/unequal, similar/dissimilar), and the judgment necessary to identify and evaluate similarities and differences. It is designed to develop the following cognitive functions: spontaneous comparative behaviour, clear, stable perception, thorough, systematic exploration, precision/accuracy, discrimination, judgement o f similarity, projection o f relationships, evaluation, judgements, conclusions, classification, planning behaviour, the use o f two or more sources o f information simultaneously, and the use o f verbal tools, discrimination, and summation. 4. Analytic Perception.  This instrument addresses the ability to analyze component parts  in order to find how they relate to each other as well as how they contribute to the overall  .19  character o f the whole they compose. It is designed to develop the following cognitive functions: comparisons, hypothetical thinking, visual transport, precision/accuracy, systematic search and work, strategy definition, labelling, projection o f relationships, and spontaneous use o f two or more sources o f information. In the second-year curriculum, students use these instruments: 5. Categorization.  The materials are designed to help students learn the underlying  principles and strategies for creating conceptual sets and categories, a vital prerequisite for higher mental processing. 6. Instructions. This instrument emphasizes the use o f language as a system for both encoding and decoding operational processes on levels o f varying complexity. Exercises focus on critiquing instruction, rewriting instructions to supply missing relevant data, and creating instructions and directions for other to follow. 7. Temporal Relations.  The materials address chronological time, biological time, and  other temporal relations. Students learn to isolate the factors involved i n evaluating or predicting outcomes (i.e., time, distance, and velocity) and to find the inter-relationships among those factors. 8. Numerical Progressions.  The instrument promotes the ability to perceive and  understand principles and formulas manifested i n numerical patterns. 9. Family Relations.  The instrument promotes understanding o f how individual roles i n  hierarchical organizations define the network o f relationships that are encountered in daily life and work. 10. Illustrations.  The materials encourage spontaneous awareness that a problem exists,  analysis o f why it exists, and projection o f cause-and-effect relationships. In the third-year curriculum, students are introduced to four instruments:  20  11. Transitive Relations and Syllogisms.  The materials are designed to foster higher-level  abstract and inferential thought. Transitive Relations focuses on drawing inferences from relationships that can be described in terms o f "greater than," "equal to," or "less than."  Syllogisms  addresses formal prepositional logic and is intended to promote inferential thinking based on local evidence. Students learn to critique analytic premises and propositions. 12. Representational  Stencil Design. The instrument requires students to analyze a  complex figure, identify its components, and then recreate the whole mentally i n colour, shape, size and orientation. 13. Orientation in Space II. The materials complement earlier instruments by extending students' understanding o f relative positions from a personal orientation to the stable, external system represented by the points o f the compass. Literature Review There have been a variety o f perspectives on the value o f F I E and I have presented them below. These differing perspectives aided me in distilling features that led to the research questions. I w i l l present them without comment at this time and return to selected opinions i n my discussions in Chapter Six. I w i l l begin with reviews and reports o f individual studies and w i l l then present reviews o f groups o f studies; they w i l l be i n chronological order. I w i l l then highlight approaches to intelligence other than Feuerstein's so that we have a context within which to consider FIE. Individual Studies After reviewing the data from Feuerstein's study (Feuerstein et al., 1980), Bradley (1983) observed that on the Primary Mental Abilities test, Project Assessment Battery and Classroom Participation Scales, validity, reliability and other psychometric data were not reported. H o w then, he questioned, can we be certain the tests measure what they are said to measure, and how  21  well they do it. Further, there was no comment in the study regarding control o f experiment-wide error rates. A failure to control for experimental error rate would result in the probability levels o f the statistical tests being much higher than the .05 level that was set. In addition, because o f the large number o f participants i n the study, the power o f the statistical tests would produce statistical significance even i f the differences among means were very small. Bradley (1983) also drew a parallel between the current interest in cognitive skills training and the optimism in the 1960s among special education circles for the diagnosis and remediation o f learning difficulties. These difficulties were related to psychoneurological processes (e.g. auditory, visual, attention or sequencing processes) and training those process was important. However, throughout the 1970s, a series o f studies concluded that the tests o f processing abilities did not relate well to cognitive or academic performance, and the methods for training processes had failed to produce desired gains in cognitive, academic or even process abilities (Hammill, 1993). Arbitman-Smith, Haywood, and Bransford (1984) demonstrated that F I E students improve greatly in their ability to master tasks that they are taught directly and to solve domain specific transfer problems and domain independent problems. They also found that F I E students benefited more from instruction unrelated to F I E , in part because they were more attentive and more motivated to engage in learning tasks. O n normative measures o f intelligence, results were encouraging but not overwhelming. Arbitman-Smith et al. (1984) concluded that one necessary condition for cognitive change seemed to be the teachers, i n that they must be trained to carry out the program effectively. They also asked how much cognitive education would be enough. According to the authors, the F I E program is designed to be taught 300 to 350 hours for a period o f 2 to 3 years. However, teachers usually average only 70 to 80 hours per year, and often cannot continue beyond the first year.  22  Another question the authors would like answered is, " W h o benefits from F I E - normal, but low academic functioning, educable mentally retarded, or the learning disabled?" Another researcher, Blagg (1991), reported on an extensive study o f F I E i n Somerset, England. Beginning in 1983, the project ran for 5 years with a formal evaluation over 3 years o f the effects o f F I E on 14- to 16-year old low-achieving adolescents, their teachers, and schools. During this period, approximately 1,000 pupils were exposed to F I E and 30 teachers and three psychologists were trained in the program. Over a 2-year period, 250 children and 30 teachers were closely monitored. The study failed to confirm any increase i n IQ scores associated with FIE. Further, there was no evidence to imply that it had a positive influence on attainments or work-study skills. There was some basis for mild optimism about positive attitudinal and behavioural change i n the pupils and clearer evidence o f benefits for the teachers who became more positive about their roles and about the potential o f low-functioning students. However, Blagg believes that, "The major stumbling block with F I E appears to be its failure to teach for transfer" (p. 135). Groups of Studies Sternberg (1984) and Sternberg and Bhana (1986) bring several concerns to our attention. In their review o f the literature they state that very little information is available on the constructs of insight and motivation, both o f which the program is intended to develop. They question how well the instruction w i l l transfer or generalize to other areas o f students' lives, partly because the extent o f transfer depends greatly on how well teachers are able to conduct the required bridging. Sternberg and Bhana (1986) also saw no evidence o f improvement in insightful, creative, or synthetic thinking abilities. Nevertheless, they state that the program would likely be useful and to be most appropriate for students o f average or below average ability at the junior high school level and for students who do not grow up with standard white, North American middle class  23  experiences. Further, they believe that the program may have the potential to improve students' standardized scores in the areas o f abstract reasoning and spatial visualization. Sternberg (1984) views the strengths o f the IE program as the following: •  It can be used for children i n a wide age range and for children o f a wide range o f ability levels and socioeconomic groups.  •  It is well liked by children and appears to be effective i n raising their intrinsic motivation and self-esteem.  •  It is well packaged and readily obtainable.  •  It appears effective i n raising children's scores on IQ tests. The weaknesses, according to Sternberg (1984), are:  •  The program requires extensive teacher training, which must be administered by a designated training authority for the duration o f the program.  •  The isolation o f the problems from any working knowledge or discipline base (such as social issues or reading) raises questions regarding the transferability o f the skills to academic and real-world intellectual tasks, especially over the long term.  •  Despite Feuerstein's dislike o f IQ tests, the program trains primarily those abilities that IQ tests tap rather than a broader spectrum o f abilities that go beyond intelligence as the tests test it. Savell, Twohig, and Rachford (1986) examined reports o f empirical research on F I E as a  method o f teaching thinking skills. They began with studies conducted i n Israel, the first o f which was conducted in the early 1970s (Feuerstein et al., 1980; Rand, Tannenbaum, & Feuerstein, 1979). The second was a follow-up study conducted 2 years later. Savell et al. concluded that the reports suggest the possibility that, at least with what Feuerstein considered to be culturally disadvantaged students, F I E is capable o f producing some lasting improvement i n the ability o f some students to do well on at least some measures o f intellectual ability. However, the similarity 24  between FIE-material and dependent-measure material raised the question as to whether or not the studies produced nothing more than mastery or near-transfer or practice effects. Further, the authors state that there is the question o f how to interpret the studies' failure to find statistically significant effects on self-concept measures. Savell et al. (1986) also reviewed reports from Venezuela, the United States, and Canada, and they arrived at the following generalizations: •  Statistically significant FIE/non F I E comparison group differences had been observed i n these countries, including Israel, in middle and low social class groups, in groups considered normal as well as groups considered culturally or educationally disadvantaged, and i n groups classified as hearing-impaired.  •  The effects most commonly reported in these studies had been those on certain standard nonverbal measures o f intelligence that are largely measures o f skill i n processing figural and spatial information. Effects on other types o f measures, such as self-esteem, impulsivity, classroom behaviour, academic achievement and course content were absent, inconsistent or difficult to interpret.  •  The effects that had been reported have been observed almost entirely with individuals who were i n elementary or secondary school (ages 12-18). A few studies used college or collegeage subjects, but generally the intervention used was too weak to provide a satisfactory test. One study suggested that F I E may be able to produce effects with individuals who are beyond adolescence.  •  Studies showing experimental/comparison group differences had a number o f things in common: at least a week o f training for F I E instructors, generally 80 hours or more o f student exposure to F I E over a 1- or 2-year period, and F I E taught in conjunction with some other subject matter o f interest and importance to the student. 25  Savell et al. (1986) ended with conclusions and questions for future research. First they perceived the need for data that: •  Goes beyond positive near-transfer results (transfer from F I E material to FIE-like material such as nonverbal tasks) to positive far-transfer results (transfer to dissimilar behaviours such as planning ability),  •  Relate to students' ability to learn when placed in standard learning situations, and  •  Address snowballing, that is, the increasing magnitude o f positive F I E results over time. Then these authors ask, in part:  •  Is F I E effectiveness related to age?  •  To what extent can F I E effects be observed in populations other than those originally studied in Israel?  •  W h y have measures o f self-esteem generally failed to show F I E effects?  •  H o w much F I E instruction on how many instruments are needed to produce the effects intended by F I E theorists?  •  H o w much training and support is needed to prepare teachers to provide effective instruction?  •  H o w does bridging take place and do regular teachers who are trained i n F I E teach their other classes differently than before?  •  What is the best way to measure F I E effects, that is, what are the best instruments for particular constructs?  •  H o w large an effect is it reasonable to look for? (pp. 402-403) In his review o f FIE, Burden (1987) found that more attention needed to be given to the  bridging aspects o f mediation, that F I E had been criticized for being too close to a teach-to-test program, and that the issue o f "optimum engaged time" had not been given much attention from  26  researchers. In addition, studies incorporating measures o f self-concept had shown little, i f any, positive effects o f FIE. Yates (1987) reviewed 18 studies conducted from 1979 to 1985. She found that there was little controlled research published regarding F I E effects and no follow-up data had been reported. Most o f the studies were small-scale ones and less than the recommended amount o f instruction was given (400 hours, according to Yates). There had been no attempt to analyze IE's impact on such factors as time-on-task, instructional support, success rate or academic learning time. A number o f studies reported modest gains in intellectual measures, and many studies reported that teachers responded enthusiastically to the IE materials. Samuels and Conte (1987) also examined the research on Instrumental Enrichment. They reported that although initial studies carried out by Feuerstein and his colleagues (for example, Feuerstein et al., 1980) indicated that Instrumental Enrichment could make significant changes on the intellectual functioning o f culturally disadvantaged children, results o f follow-up studies i n North America were less encouraging. There were, however, problems with the research. First, students often received less than the recommended amount o f instruction, which, according to Samuels and Conte, should be no less than 3 hours a week with a total m i n i m u m o f 200 hours. Second, studies did not monitor teacher behaviour to assess how well the teachers had adopted a mediated teaching style. Regarding the evaluative instruments used i n the research, the measures used were standardized tests o f intelligence, achievement, self-esteem and so on. Samuels and Conte (1987) wrote that the tests had been criticized for a lack o f sensitivity to the processes and strategies being taught. In their own research with learning-disabled students at a vocational school, the authors found that although results first appeared to be disappointing, there was an effect due to differential attrition over the term o f the study. Poor students were encouraged to remain in school  27  and higher functioning students were encouraged to transfer to other schools offering academic programmes. Frisby and Braden (1992) examined studies by Perry (1986), Bachor (1988), Haywood, Towery-Woolsey, Arbitman-Smith and Aldridge (1988), and a study by Shayer and Beasley (1987) on data from three research programs. Shayer and Beasley gave effect sizes for results reported by Feuerstein (1979) and Haywood, Arbitman-Smith, Bransford, Towery, I.L. Hannel, and M . V . Hannel (1982) and from their own study. In the first study, the F I E program did not result i n improvement o f school-related behaviours. The second succeeded on near-transfer tasks, and the third, involving hearing-impaired students, also showed significant gains mainly on measures o f nonverbal ability and spatial relations. In the Shayer and Beasley study, most achievement subtest effect sizes were nonsignificant. Further, effect sizes that did achieve significance translated into a few subtest points difference, carrying little practical significance. Frisby and Braden (1992) also wrote a detailed critique o f Feuerstein's dynamic approach to assessment. In the critique, they challenged Feuerstein's use o f the term "cognitive modifiability." They refer to Eysenck (1988), stating that arguments over the extent to which intelligence is modifiable w i l l become hopelessly muddled unless the level o f intelligence that is referred to is carefully clarified. Frisby and Braden believe that this clarification requires a distinction between biological intelligence, psychometric intelligence and social intelligence. According to Frisby and Braden, biological intelligence refers to individual differences in the structure, physiology and biochemistry o f the human brain that are determined by one's genetic code. Psychometric intelligence refers to a person's general level o f functioning on objective, standardized, intelligence tests. Social intelligence is meant by the authors to refer to the ability to function and adapt within social or everyday situations according to the accepted standards o f  28  one's culture. The authors believe that judgments o f social intelligence vary with the persons making the judgment and the context within which judgments are made. Frisby and Braden (1992) claim that, because many noncognitive factors enter into the determination o f social intelligence, the consideration o f noncognitive factors helps to explain why social intelligence is highly modifiable. They conclude that the behaviour and attitude changes that occur as a result o f exposure to F I E training fall under the domain o f what they define as social intelligence. They state that modifiability in this situation should not be confused with cognitive test score changes or qualitative shifts i n thinking typically associated with biological maturation. Frisby and Braden (1992) also claim that attempts to draw links between ethnic/social class group differences i n cognitive performance and associated differences in childrearing styles cannot prove or disprove an ad hoc M L E hypothesis, because such accounts do not control for genetic kinship, environmental or school factors. (Feuerstein would likely claim that these factors are not very relevant because they are "distal" causes o f intellectual performance. See page 12 i n Chapter One). There are those, however, who would disagree with Frisby and Braden. The theory has become attractive to minority educators in the United States and Canada as an explanation for the low academic performance o f poor minority and aboriginal children relative to more advantaged students (for example, Wilgosh & Mulcahy, 1993). Emerson (1991) presents a case for the use o f M L E in the education o f Native Americans. H e writes that, Increased cognitive efficiency and knowledge o f native cultural information can prove to be very beneficial for contemporary native cultural survival and development. The tribal elder who sat down and mediated both the meaning behind an object and the process o f its development allowed both culture and cognition to flourish in the minds o f its youth, making them more susceptible to change and to learning, (p. 155)  29  Intelligence In this section I w i l l highlight other theoretical approaches to the way we discuss intelligence i n order to provide a context within which F I E can be considered. In a report to the American Psychological Association, Neisser, Boodoo, Bouchard, B o y k i n , Brody, Ceci, Halpern, Loehlin, Perloff, Sternberg, and Urbina (1996) reviewed the literature on various concepts o f intelligence. They attempted to "prepare a dispassionate survey o f the state o f the art" (p. 78) and review four concepts o f intelligence: the psychometric approach, multiple forms o f intelligence, biological approaches, and developmental progressions. Their concepts o f psychometric and biological intelligence appear to be similar to those o f Frisby and Braden. In their description o f developmental progressions o f intelligence, Neisser et al. included the theories o f Jean Piaget and L e v Vygotsky. D a v i d Wechsler originally defined intelligence as the capacity o f the individual to act purposefully, to think rationally, and to deal effectively with his environment" (Wechsler, 1944, p. 3). He believed that intelligence was not only a global entity but also an aggregate o f specific abilities. Later, he explored "nonintellective" factors o f intelligence, including the abilities to perceive and respond to social, moral and aesthetic values (The Psychological Corporation, 1997; Wechsler, 1975). These intellectual factors appear to be similar to the concepts o f social intelligence as described by Frisby and Braden, and to Feuerstein's concepts o f cognitive modifiability and mediated learning experiences. According to Sternberg (1994), "we need to move away from a conception o f intelligence as constituting a fixed set o f abilities, regardless o f the number, and toward a conception o f intelligence as involving capitalization on strengths and compensation for and remediation o f weaknesses" (p. 563). He believes that we can teach intelligence to at least some degree, but cannot effect radical changes at this time (Sternberg, 1996).  30  Perhaps the discussion o f whether or not intelligence can be modified would be more fruitful i f we consider Haywood's "transactional" theory. In this model, "individual differences in intellectual development and expression are seen to be products o f genetic endowment engaged in a series o f 'transactions' with environmental circumstances, and with a person-characteristic trajectory o f development for each individual" (Haywood, Tzuriel, and Vaught, 1992, p. 45). Haywood believes that intelligence is only modestly modifiable, but cognitive processes (or Piagetian "cognitive structures"), that is, durable and generalizable modes o f thinking, are highly modifiable. He contends that components o f intelligence include several factors all related to socalled "pure" ability variables, but components o f cognitive processes may include both cognitive "structures," and motivational/attitudinal/affective variables. This view appears to be consistent with the views o f Wechsler (1944) and Sternberg (1994), and with Feuerstein's theory o f Structural Cognitive Modifiability. So, when we discuss the effectiveness o f cognitive modification, we can think, not o f intelligence, but o f modes o f thinking and behaving.  Summary of Evaluations The evaluations o f empirical research, above, contained not only negative but positive aspects. However, there were many questions and areas noted for further investigation. Several authors were concerned about an appropriate amount o f required teacher training, as well as about teachers' skills with a mediational teaching style and with bridging. They were also concerned about the length o f time that was spent teaching Instrumental Enrichment, often far less than a recommended amount. There were questions about the generalizability o f IE, that is, whether positive effects were only on near-transfer tasks or when students were tested with instruments similar to IE instruments. There also seemed to be little success on measures in the affective domain.  31  Reviewers wanted information on what types o f populations benefited most from IE and at what age. They wondered how large an effect size would be considered important. They also recommended that large follow-up studies should be conducted, including evaluations o f students who left the IE program before the end o f a study. Factors such as time-on-task, teacher support and which instruments were most important for which behaviour/learning changes also needed research. Finally, many reviewers noted the need for more rigorously controlled studies. The review o f the literature has, therefore, highlighted gaps in our understanding o f Instrumental Enrichment. There appear to be three areas for consideration regarding the effectiveness o f IE: first, there are factors which have an impact on the curriculum; that is, the amount o f teacher training required, the skill level o f the teachers, the time spent teaching I E , and the type and number o f instruments used. The second area concerns factors that have an impact on the results o f IE, that is, whether IE transfers to the cognitive and affective domains. The third area for study is related to the populations that IE effects, that is, whether different groups benefit from the instruction. The research questions, arising from the literature review, are: 1. What is the optimal amount o f teacher training required i n order to best teach Instrumental Enrichment? 2. H o w skilled were teachers with a mediational teaching style and with bridging, and did this make a difference to student outcomes? 3. What is the optimal length o f time that should be spent teaching Instrumental Enrichment for best results? 4. H o w many instruments are needed to produce optimal results?  32  5. H o w far does Instrumental Enrichment transfer? That is, specifically: 5a) A r e there significant effect sizes for cognitive tasks such as full scale reasoning ability, and verbal and nonverbal reasoning ability for students who have participated i n Instrumental Enrichment programs? 5b) A r e there significant effects for measures o f academic achievement and in the affective domain on measures o f self-esteem, self-confidence, self-concept, motivation to learn, behaviour and attitudes? 5c) A r e significant effect sizes restricted to tasks similar to those in the F I E curriculum, for example, visual-spatial tasks, visual-motor tasks and dynamic assessment with the Learning Potential Assessment Device? 6. What types o f populations benefited from IE instruction? Populations that have been studied include those related to gender, ethnicity, socioeconomic status, age, physical handicaps, intellectual ability, learning disabilities, and emotional and behavioural handicaps. Although these are questions that could be addressed, because o f the nature o f the data analysis (described in Chapter Four), results o f this meta-analysis w i l l be discussed i n terms o f Question 5, that is, i f results were significant for measures o f near and far transfer.  Importance of the Study I believe this exploratory investigation is important because so many o f the results o f individual studies have been inconclusive or contradictory and a meta-analysis is a rigorous method o f combining all o f their results. Information from variables I am able to analyze may provide insight into the chosen research question, above. A meta-analysis is also important because, as Bradley (1983) and Savell et al. (1986) note, it is necessary to calculate more than statistical significance between groups. In order to make effective educational decisions we  33  should know how large the difference can be between the groups. It may be that i n certain circumstances, the amount o f improvement is not enough to justify expenditures; i n others, the difference could be enough to justify pursuing a vigorous complete or partial program. In Chapter Three I w i l l describe meta-analysis in more detail, that is, why it is appropriate and how it is performed.  34  Chapter Three  Literature Review of Meta-Analysis This chapter w i l l describe the purpose and methodology o f meta-analysis. I w i l l begin with definitions o f what meta-analysis is and why it is a useful methodological tool. I w i l l then present a critique from the literature about meta-analysis. The critique offers discussions on theory development, Type II errors, validity, reliability, homogeneity, generalizability and bias. The bias section contains discussions on the "file-drawer" problem, quality o f studies, multiple publications, sample size, studies with more than one statistic, and the unbiased estimate o f effect size. I w i l l then describe the usual steps followed i n performing a meta-analysis. These include the literature search, developing inclusion-exclusion criteria, data collection and the statistical analysis. There are many decisions to be made as the meta-analysis progresses and many o f these w i l l be made as an iterative process as studies are collected, read and recorded. The decisions and their rationale w i l l be described i n Chapter four.  Definitions Meta-analysis was developed by Gene Glass (1976, 1977). H e defined meta-analysis as the analysis o f analyses. According to Glass, a meta-analysis is ... the statistical analysis o f a large collection o f analysis results from individual studies for the purpose o f integrating the findings. It connotes a rigorous alternative to the casual, narrative discussions o f research studies that typify our attempts to make sense o f the rapidly expanding research literature. (1976, p. 3) K u l i k and K u l i k (1989) state that a meta-analysis covers review results from large numbers of studies found i n objective searches o f a research literature. They describe it as an application o f statistical tools to summary statistics, not raw data. The observations, according to Cooper (1998) and K u l i k and K u l i k , are means, standard deviations, odds ratios and results from t and F  35  statistical tests. Correlation coefficients can also be used. A meta-analysis may focus on size o f treatment effects, not just statistical significance. According to Cohen (1988), the concept o f treatment effects is important because i n most educational studies statistical tests do not provide any insight into the strength o f the relationship or effect o f interest, so it is desirable to provide indexes o f effect size. Without intending any necessary implication o f causality, it is convenient to use the phrase "effect size" to mean "the degree to which the phenomenon is present i n the population," or "the degree to which the null hypothesis is false." Whatever the manner o f representation o f a phenomenon i n a particular research i n the present treatment, the null hypothesis always means that the effect size is zero. (pp. 9-10)  Usefulness Combining Study Results According to L'Abbe, Detsky, and O'Rourke (1987), meta-analysis is especially useful when results from several studies are inconclusive with regard to magnitude or direction o f effect, when sample sizes are individually too small to detect an effect and label it statistically significant, or when a large trial is too costly and time-consuming to perform. Although meta-analysis has often been applied to combine the results o f randomized trials, there are many topics for which randomized trials are impossible, and meta-analyses o f nonexperimental studies are also common. For nonexperimental studies, the method is also most useful when there are many studies with l o w statistical power (Petitti, 1994). M a n y o f these situations arise with studies o f Feuerstein's Instrumental Enrichment. A l o n g with the problem o f effect sizes generally not being reported i n F I E research, these are the reasons why the methodology o f meta-analysis has been chosen. M a n y F I E studies have inconclusive or contradictory results, sample sizes are often small, most, i f not all, studies are nonexperimental, and many have l o w statistical power.  36  Critique Theory Development For several reasons, researchers believe that meta-analysis can contribute to our development o f theories. Researchers can show that, although there may be statistical significance associated with a relationship, the effect size can be so small that the relationship is relatively unimportant and so does not support a theory. Meta-analysis can lead to a deeper analysis o f new or additional variables o f theoretical interest. Researchers can compare strengths o f theoretical relations that have not been compared before and offer alternative explanations for phenomena. Meta-analysis can also show us where the gaps are in research (Asher, 1990; H a l l , Rosenthal, Tickle-Degnen, & Mosteller, 1994). This may be the case with Feuerstein's Instrumental Enrichment, where, for example, few studies have been found relating to teacher training and effectiveness. It should be noted that a meta-analysis could also give a false positive effect because o f the large number o f data points. According to L ' A b b e et al., (1987) consistent findings i n different populations under different circumstances that may be shown through meta-analysis may strengthen the case for a causal association. Hall et al. (1994) do not believe, however, that we can confidently assign causality to a relationship based on meta-analytic review evidence alone. They believe that the boundaries to how much we can learn about causation with research synthesis are defined by the underlying primary studies. If most o f the primary studies are experimental and have high internal validity with, for example, randomly assigned subjects to conditions, these authors believe that then a synthesist might attempt to address questions o f causation. If most o f the primary studies have low internal validity with, for example, non-experimental treatments on pre-existing groups (as is the case with FIE), a synthesist can ask whether two variables covary, but cannot confidently address questions o f causation. Procedures that may suggest causal inference, like random  37  assignment o f subjects to conditions and control over the intervention, are available to the primary researcher. The authors acknowledge that meta-analyses may, however, provide guidance for the directions for new primary research. Hunter and Schmidt (1990), and Schmidt (1992), on the other hand, believe that metaanalysis can have more o f a role in theory development. Schmidt states that theories are causal explanations o f the processes that take place in a phenomenon. In the behavioural and social sciences, the methods o f path analysis can be used to test causal theories when the data meet the assumptions o f the method. The relationships revealed by meta-analysis, that is, the empirical building blocks for theory, can be used in path analysis to test causal theories whether or not the defined relationships are observational or experimental. Hunter and Schmidt state that it is necessary only to transform effect size values into correlation coefficients. It is the purpose o f this investigation to explore relationships and evaluate Feuerstein's Instrumental Enrichment, not to develop theories. Other researchers, however, may choose to use the results to do so.  Type II Errors According to Asher (1990), Hunter and Schmidt (1990), and Schmidt (1992), combining studies in meta-analyses is important because Type II errors can be predominant i n primary studies. This is because individual studies may not have enough data points to increase the power o f the statistical tests. Rosenthal (1991) recommends that primary researchers always report both an estimate o f effect size and a test o f significance or confidence interval to protect against inferential invalidity with Type I and Type II errors. He contends there is little doubt that i n the social and behavioural sciences Type II errors are far more likely that Type I errors. He believes that the frequency o f Type II errors can be reduced by attention to the magnitude o f the estimated effect size. If that  38  estimate is large and there is a nonsignificant result, researchers should not assume that the means o f two groups are the same. Only i f the pooled data o f many replications result in both a very small effect size on the average and a combined test o f significance that does not reach a desired alpha level are researchers justified in concluding that no nontrivial relationship exists among the variables. A meta-analysis is the best way to investigate the data.  Validity It has been claimed by Eysenck (1978), Gallo (1978), and Presby (1978) among others that logical conclusions cannot be drawn by comparing and aggregating studies that include different measuring techniques, definitions o f variables, and subjects because they are too dissimilar. According to W o l f (1986), this criticism has been referred to as the "apples and oranges problem," in which it is argued that diversity makes comparisons inappropriate. However, W o l f believes that this issue may be dealt with empirically by coding the characteristics for each study and statistically testing whether these differences are related to the meta-analytic results. L A b b e et al. (1987) agree, stating that meta-analysis offers the opportunity to explore the effects o f these differences through comparison o f the magnitude o f the treatment effect across studies, and i n the overall meta-analysis result. Through sensitivity analysis, studies with different design characteristics can be included or excluded from the meta-analysis, and the effect o f combining results from potentially dissimilar studies may be explored. Further, Cronbach and Meehl (1955) advocated the multimethod-multitrait approach o f analyzing variable differences to develop construct validity. In this method, there should be several types, approaches and methods o f assessing treatment and criteria variables, and these should be used for several types o f populations. Although this is seldom done in any one research study, Asher (1990) believes that general construct validity can be established quite powerfully  39  with meta-analysis as the researcher considers the homogeneity o f results using different types o f treatments, criteria variables, populations, experimenters and methods. Regarding the generalizability o f an effect, H a l l et al. (1994) believe that research synthesis is the ideal method. They contend that meta-analytic techniques allow for empirical assessment o f external validity to an extent rarely available to a single primary study. In primary research, confidence i n the external validity o f results is based solely on the judgment that the populations o f subjects, situations and procedures that are o f interest are adequately represented by the study. They believe that i f the phenomenon is conceptually broad and therefore should be demonstrated over a wide variety o f contexts, then studies that vary extensively i n subjects, situations and procedures may be appropriate for inclusion.  Reliability When conducting a meta-analysis, it is necessary to consider the reliability, or consistency, of locating research results. W o l f (1986) suggests that we must ask how likely it is for independent analysts to locate and include the same studies, and how comprehensive the collection of studies is. Another important issue is the degree o f reliability i n coding the features o f studies. There are guidelines to enhance intercoder reliability (for example, Orwin, 1994, and Petitti, 1994). However, these caveats assume that more than one person w i l l be working on the metaanalysis, and i n this case there w i l l be only one, myself. Another reliability issue is the degree o f consistency in the calculating and recording o f the significance levels and effect size estimates o f studies. M a n y o f these outcome results are not reported directly i n the primary studies, and W o l f (1986) believes that the possibility o f errors i n their calculations and coding needs to be minimized.  40  Homogeneity In order to quantitatively synthesize all studies in one meta-analysis, it is assumed that each study provides a sample estimate o f the size o f effect that is representative o f the population effect size. If a series o f independent studies provide a common, or homogeneous, estimate o f the population effect size, then it is more likely that the various studies are testing the same hypothesis. The underlying assumption i n combining individual study results to arrive at a summary measure is that their differences are due to chance alone (sampling variation), and, therefore, all study results are homogeneous, that is, that they reflect the same "true" effect. In other words, when results are combined, random error is reduced and many observations are better than one (L'Abbe et al., 1987; Wolf, 1986). Tests o f homogeneity among the estimates o f effect size are often performed before researchers examine moderator variables. The tests measure whether a group o f effects is more variable than one would expect based on sampling variation. A significant test means that greater variability than expected by sampling error is present (Hall et al., 1994). If the estimates are heterogeneous, then the question arises as to whether each study is testing the same hypothesis. Heterogeneity provides a warning that it may not be appropriate to combine and synthesize all the study results i n one meta-analysis. The investigator would need to consider whether to conduct separate meta-analytic syntheses for different subsets o f studies, each o f which represent homogeneous results (Cooper, 1998; Wolf, 1986). The power o f statistical tests o f homogeneity is low, and the failure to reject the hypothesis that the studies are homogeneous does not prove that the studies are measuring the same quantity (Petitti, 1994). However, the analyst can use a higher alpha level, for example, .10 instead o f the usual .05, to decrease the risk o f having a Type II error. The issue o f homogeneity/heterogeneity w i l l be discussed in more detail below.  41  Sources of Variation and Generalizability The decision about whether or not to use a fixed or random effects model to analyze the data depends upon the degree o f homogeneity o f the individual studies. According to Petitti (1994), in all the methods based on the assumption o f fixed effects, the variance component o f the meta-analytic estimate o f pooled effect size is composed only o f terms for the within-study variance o f each component study. The assumption o f the random-effects model that studies are a random sample from some population o f studies makes it necessary to include a between-study as well as a within-study component o f variation i n estimation o f effect size and statistical significance. Because the random-effects model incorporates a between-study component o f variance, it w i l l be more "conservative" than an analysis o f the same data based on a method that assumes fixed effects. That is, an analysis based on a random-effects model w i l l generally yield a wider confidence interval and w i l l be less likely to declare a significant difference than an analysis based on fixed effects. A s the between-study variance becomes large, that is, as heterogeneity increases, the between-study variance term w i l l dominate the weights assigned to the study using the randomeffects model, and large and small studies w i l l tend to be weighted equally. In this situation, the results o f an analysis based on the fixed-effects model, which weights studies according to their sample size, and the random-effects model may differ considerably. When there is not much heterogeneity, the fixed-effects and the random-effects models w i l l both tend to weight studies primarily according to sample size and they w i l l yield results that are similar. Petitti (1994) contends that there is no empirical basis for preferring the fixed-effects model over the random-effects model or vice versa. The choice o f fixed-effects model and a random-effects model is secondary to the examination o f the reasons for lack o f homogeneity. If studies are homogeneous, then the choice between the fixed-effects model and the random-effects  42  model is unimportant, as the models w i l l yield results that are similar. It has been argued that one should not pool disparate study results at all. Rather, the results should be reported and modeled, or the reasons for the lack o f homogeneity should be examined. Petitti states that the use o f the random-effects model is not considered to be a defensible solution to the problem o f heterogeneity. The desire to be conservative is a reason to use the random-effects model, but only in the absence o f heterogeneity. Cooper (1998), on the other hand, writes, The question we must ask is whether the effect sizes in a data set are affected by a large number o f these uncontrollable influences, such as differences i n teachers, schools, family structures, and so on. If the answer is "yes, they probably are," then the meta-analyst chooses a statistical model that takes these additional sources o f random variance i n effect sizes into account. If the answer is "no, probably not," then random variance in effect sizes is ignored (or more accurately, set to zero) and a fixed-effects statistical model is used. (pp. 150-151) According to Petitti (1994), i n an analysis based on a fixed-effects model, inference is conditional on the studies actually done. In an analysis based on a random-effects model, inference is based on the assumption that the studies are a random sample o f some hypothetical population o f studies. The statistical methods used to combine study results when fixed effects are assumed differ from the methods used when random effects are assumed. However, it has been shown that differences in the results o f meta-analysis based on fixedeffects and random-effects models arise only when the study results are not homogenous. There are strong opinions about the appropriateness o f both models. Several statisticians have expressed a preference for the fixed-effects approach, while others favour the random-effects approach (Cooper, 1998; Petitti, 1994). It is generally agreed, however, that  43  the questions addressed by analysis based on the fixed-effects model and based on the randomeffects model are different. According to Petitti, the random-effects assumption means that the analysis addresses the question, " W i l l the treatment produce benefit 'on average'?" The model is appropriate i f the question is whether the treatment, or the risk factor, will have an effect. The fixed-effects assumption means that the analysis addresses the question, " D i d the treatment produce benefit on average in the studies on hand?" If the question is whether the treatment has caused an effect in the studies that have been done, then the fixed-effects model is appropriate. Hedges (1994) uses the term "universe" to denote the hypothetical collection o f studies that could be conducted in principle and about which the analyst wishes to generalize. H e uses the term study "sample" to denote the collection o f studies that are used in the review and that provide the effect size data used in the research synthesis. In the fixed effects, or conditional, model, the universe to which generalizations are made consists o f collections o f studies identical to those i n the study sample except for the particular people, or primary sampling units, that appear i n the studies. Thus, the studies in the universe differ from those in the study sample only as a result o f the sampling o f people into the groups o f the studies. The model is called the conditional model because it can be conceived o f as a model that is conditional on the studies observed. The conditional model in research synthesis is i n the same spirit as the usual regression model and fixed effects analysis o f variance i n primary research. In conditional models inferences to other cases are, i n the strictest sense, limited to cases i n which the collection o f values o f the predictor variables is represented i n the sample.  44  However, Hedges believes that inferences may be generalized to all studies i n the universe i f the studies sampled are sufficiently similar to those in the universe. In the random effects, or unconditional, model, the study sample is presumed to be literally a sample from a hypothetical collection, or population, o f studies. The universe to which generalizations are made consists o f a population o f studies from which the study sample is drawn. Studies i n this universe differ from those in the study sample along two dimensions. First, the studies differ from one another in study characteristics and in effect size parameter because the effect size parameter has a distribution. The generalization is not, as it was in the fixed effects case, to a universe consisting o f collections o f studies with corresponding members o f the collections having identical characteristics and effect size parameters. Instead the studies in the study sample (and their effect size parameters) differ from those i n the universe by as much as might be expected as a consequence o f drawing a sample from a population. Second, in addition to differences i n study characteristics and effect size parameters, the studies i n the study sample also differ from those i n the universe as a consequence o f the sampling o f people into the groups o f the study. This results i n variation o f observed effect sizes about their respective effect size parameters. This model is called the unconditional model because, unlike the fixed effects model, it does not condition, or hold fixed, the characteristics o f studies that might be related to the effect size parameter. The random effects model in research synthesis is i n the same spirit as the correlation model or the random effects analysis o f variance i n primary research. There are two sources o f uncertainty in estimates and tests in random effects analyses: one due to the sample o f the treatment effects themselves and the other due to the sampling o f individuals (in particular, outcome scores) into each treatment.  45  Inferences are not limited to cases with predictor variables represented i n the sample. Instead, the inferences apply to the universe o f studies from which the study sample was obtained. Because the universe contains studies that differ in their characteristics, and those differences find their way into the study sample by the process o f random selection, generalizations to the universe pertain to studies that are not identical to those in the study sample. B y utilizing a random effects model o f generalization, the analyst does not have to ask, " H o w similar is similar enough?" as with the fixed effects model. Instead, the question is, "Is this new study part o f the universe from which the study sample was obtained?" In metaanalysis the universe is usually rather ambiguously specified, and consequently the ambiguity in generalization based on random effects models is that it is difficult to know precisely what the universe is. In contrast, the universe is more clear in fixed effects models, but the ambiguity arises in deciding i f a new study might be similar enough to the studies already contained i n the study sample. This can be tested with, for example, a test o f homogeneity.  Bias The File-Drawer Problem. L ' A b b e et al. (1987) and W o l f (1986) write that published research is biased i n favour o f significant findings because nonsignificant findings are rarely published, and this leads to biased meta-analysis results. Rosenthal (1979, 1984) also believes it is unlikely that a literature review w i l l uncover every study o f a hypothesis that has been conducted. He called this the "file-drawer problem" because o f the tendency for studies supporting the null hypothesis o f no significant results to likely be buried away in file drawers. Rosenthal claims that this may enhance the likelihood o f a Type I publication bias error in finding more positive results than would be the case i f all studies could be located and included in an analysis. Rosenthal suggested that we  46  address this problem by calculating the number o f studies confirming the null hypotheses that would be needed to reverse a conclusion that a significant relationship exists. This has been called the " F a i l Safe N " (Cooper, 1998; Wolf, 1986). Petitti (1994), however, recommends against using Rosenthal's method. She maintains that the statistical theory that would allow this to be done is not well developed, and the methods that claim to allow these calculations to be made are all ad hoc. W o l f (1986) believes that the analyst can overcome the problem by reviewing results in books, dissertations, unpublished papers presented at professional meetings, and so on. Another proposed solution to the problem o f publication bias is the use o f a Funnel Plot, or Graph (Greenhouse & Iyengar, 1994; Petiti, 1994). W i t h this procedure, described by Light and Pillemer (1984), the effect measure is plotted on the horizontal axis and the sample size on the vertical axis. Since small studies usually show more variability among the effect sizes than larger studies, and there w i l l typically be fewer o f the latter, the plot should look like a funnel that is viewed with the large opening down and the tip pointed up and centred on the true effect size. When there is publication bias against, for example, studies showing small effect sizes, the left side o f the plot w i l l be distorted or missing. Petiti reports that although the sensitivity o f funnel plots as a method for detecting the existence o f publication bias has not been systematically assessed, bias should be suspected when a plot is distorted.  Quality. The internal validity o f meta-analysis depends on complete and accurately reported information i n published articles. Surveys o f the literature have shown that important information is often unreported. This is not a weakness o f meta-analysis itself, but points to the need for more standardized and comprehensive reporting in the clinical literature (L'Abbe et al., 1987).  47  In addition, there is the belief that meta-analysis disregards the quality o f studies, in that only "good" studies should be aggregated (Eysenck, 1978). Slavin (1986) believes that only the best evidence should be used in forming a judgment about effects i n an area o f study. H e compares the research reviewer to a judge in a court o f law. The judge would use the best possible evidence in reaching a decision, and the researcher should do the same. According to K u l i k and K u l i k (1989), Slavin's study-inclusion criteria unnecessarily reduces the number o f studies available for analysis and with reduced pools o f studies he and his colleagues were able to carry out only the most rudimentary statistical analyses. K u l i k and K u l i k also concluded that Slavin's best-evidence approach did not provide sufficient safeguards against personal biases o f the analyst. Glass, M c G a w , and Smith (1981) and K u l i k and K u l i k (1989) believe that meta-analysts should be tolerant o f possible methodological flaws. They include studies that vary i n quality and source and those which are true experiments and quasi-experiments. They believe that metaanalyses w i l l provide good answers only i f analysts examine studies that vary i n their features. L A b b e et al. (1987) and W o l f (1986) write that meta-analysis can incorporate the assessment o f study quality by varying inclusion and exclusion criteria, using a weighting scheme when pooling results, or by coding the quality o f the design employed in each study and examining whether the results differ for poorly and well designed studies. Other suggestions for examining validity include coding for the degree o f experimenter blindness, randomization, sample size, controls for recording errors, type o f dependent variable, and, as discussed above, publication bias (Wolf, 1986).  Multiple Publications. Information from the same study should contribute only once to the summary estimate o f effect. Failure to exclude multiple reports o f the same study has the potential to cause bias i n the  48  summary estimate o f effect (Petitti, 1994). Petitti recommends that the criteria for choosing among results from multiple publications be defined.  Sample Size. Some studies i n a meta-analysis may be based on very small or unrepresentative samples o f subjects, while others may use more carefully designed randomized control group designs with large numbers o f subjects. To give these studies equal weight could lead to the less representative studies contributing just as much weight to the results o f the analysis as the more well-designed studies (Wolf, 1986). Hedges and O l k i n (1985) note that the variance o f the estimator depends on the sample size, so that estimates from studies with larger sample sizes are more precise than those from studies with smaller sample sizes. They suggest that when the primary studies do not have a common sample size, there should be some weighting and they offer methods for combinations o f linearly equatable, that is, homogeneous, measures. They demonstrate that the weights that minimize the variance o f the overall effect sizes are inversely proportional to the variance o f the estimates in each study (pp. 108-114).  Studies With More Than One Statistic. Multiple results from the same study are often used. This may bias or invalidate the metaanalysis and make the results appear more reliable than they really are because these results are not independent. To overcome the difficulty, some analysts choose to perform separate analyses for each different outcome, others choose to lump them into the same analysis, some choose to limit themselves to a fixed number o f results from each study, while others take the average o f all results from the same study (Cooper, 1998; Wolf, 1986). K u l i k and K u l i k (1989) argue that using multiple results, or what they term inflation o f sample sizes, is a major problem i n statistical analyses o f effect sizes. Many meta-analysts attempt to "dredge" multiple estimates from the studies. This w i l l have to be addressed with Feuerstein's work where it is common that many  49  disparate measures are used i n individual studies to test the significance o f the effects o f Instrumental Enrichment. Several approaches have been used for this problem o f multiple dependent variables. K u l i k and K u l i k (1989) report that they seldom code more than one effect size from one study for any given analysis, but they often carry out more than one separate analysis in a report. For example, i n an analysis o f the effects o f elementary computer-based education, they conducted separate sets o f analyses o f several subgroups such as sex, grade-level, subject matter, attitude, and so on. However, i n each analysis, the number o f effect estimates was equal to the number o f studies with the relevant data. W o l f (1986) believes that conducting separate meta-analyses for different classes o f outcome dependent variables addressing different hypotheses o f interest is a practical solution to the "apples and oranges" criticism o f meta-analysis. He states that the approach enhances the conceptual clarity and interpretation o f the results as each category o f outcome or dependent variable is examined in a separate meta-analytic synthesis. In this way, he says, apples are treated as apples, and oranges as oranges.  Unbiased Estimate of Effect Size. Focusing on Cohen's (1988) effect-size estimator d, which is calculated by subtracting the average score o f the untreated group from the average score o f the treated group and then dividing the difference by the pooled estimated standard deviation for the two groups, Hedges (1981, 1982) developed the sampling distribution o f d and showed that it is a slightly biased estimate o f an underlying population effect size. He proposed a correction for d that removed its bias, but other meta-analysts soon reported that use o f this correction had at most a trivial effect on the results. K u l i k and K u l i k (1989) found on one meta-analysis that uncorrected and corrected effect sizes correlated .999, and in most cases agreed to two decimal places. M a n y meta-analysts no longer  50  use this correction and K u l i k and K u l i k reported that Hedges (1986) conceded that modifications are needed before his formulas for effect size and standard errors o f effect sizes can be used i n meta-analysis.  Procedure There are four basic steps in a meta-analysis after the hypotheses o f interest are stated. First, studies with relevant data are identified. Second, eligibility criteria for inclusion and exclusion o f the studies are defined. Third, data are abstracted. Fourth, the abstracted data are analyzed statistically (Petitti, 1994)  Literature Search According to Petitti (1994), a critical feature o f the proper application o f the method o f meta-analysis is development o f systematic, explicit procedures for identifying studies with relevant data. The systematic, explicit nature o f the procedures for study identification distinguishes meta-analysis from qualitative literature review. In being systematic, the procedures reduce bias. In being explicit, the procedures help to ensure reproducibility and reliability. Petitti contends that no matter how sophisticated the statistical techniques used to aggregate data from studies, a review does not qualify as meta-analysis unless the procedures to identify studies are both systematic and explicit. Following Cooper (1998) and Petitti (1994), identification o f studies w i l l include a search of personal reference files and a computerized search o f databases such as the Educational Resources Information Centre (ERIC), Psychological Abstracts, Dissertation Abstracts and the Internet for published and unpublished reports, dissertations, and other fugitive literature. The title and abstract o f studies identified i n the computerized search w i l l be scanned to exclude any that are clearly irrelevant. The full text o f the remaining articles w i l l be retrieved, and each paper w i l l be read to determine whether it contains information about an experimental or observational  51  study o f FIE. The reference lists o f these articles w i l l be reviewed to identify citations to other studies o f FIE, and works that were not identified i n the computerized literature search w i l l be retrieved and reviewed for presence o f relevant information. Reference lists o f review articles w i l l also be reviewed to check for completeness o f the assembled list o f relevant publications. Research into F I E has been conducted in many countries and whenever possible, pertinent data published i n languages other than English w i l l be identified. Individuals connected to research into F I E w i l l also be personally contacted. Finally, a decision w i l l be made whether or not to use Rosenthal's Fail-Safe N or other methods that address the file-drawer problem.  Inclusion-exclusion criteria The goals o f defining eligibility criteria are to ensure reproducibility o f the analysis and to minimize bias in selection o f studies for the analysis. Another analyst faced with the same body o f literature applying the same eligibility criteria should choose the same set o f studies. The studies chosen for the meta-analysis should be as unbiased as possible with respect to their results and their conclusions (Petitti, 1994). The rationale for choosing the criteria w i l l be stated. In practice it may be necessary to revise broad inclusion criteria after reviewing all the studies found with the literature search. Some o f the variables on which inclusion criteria can be based are the study design, sample size, type o f experimental and control therapies, whether or not the study is published, and the outcomes o f interest (L'Abbe et al., 1987; Petitti, 1994). The inclusive dates o f publication, presentation, or conduct o f studies eligible for the analysis w i l l be given. The meta-analysis w i l l be as up-to-date as possible, and the cut-off date for identification o f eligible studies w i l l be specified so that material published after the cut-off date w i l l not be assumed to have been missed in the literature search. In addition, the eligibility o f studies whose results are not available in English w i l l be discussed. The criteria  52  for choosing among results o f multiple publications from the same study population w i l l also be defined. A n y restrictions on eligibility due to sample size w i l l be stated. It may be necessary to exclude very small or highly unbalanced studies. Follow-up studies can determine i f the effects o f FIE last over time. Eligibility or ineligibility based on the similarity of treatments or exposure w i l l be considered. In the case o f this analysis, the treatment is always instruction in Instrumental Enrichment. The amount o f instruction and the number o f instruments used varies, however, and w i l l have to be considered. The eligibility or ineligibility o f incomplete and unpublished reports w i l l be addressed. These could include an unpublished report from a meeting, an abstract which may or may not have been published, a letter or a brief report (Petitti, 1994). Studies can be judged for quality, selected according to a pre-determined level, or quality scores can be used to weight the individual study results in the pooling process and they can be coded accordingly (L'Abbe et al., 1987). A g a i n following Petitti, i f a study is judged to be ineligible the results w i l l be recorded, and a log o f ineligible studies w i l l be maintained. Studies that are directly pertinent but are not included in the meta-analysis w i l l be cited i n the published report on the meta-analysis, and reasons for rejecting them w i l l be presented.  Data Collection Following Cooper (1998), K u l i k and K u l i k (1989), Rosenthal (1991), and Stock (1994), several different types o f study features w i l l be recorded. These can include the features o f the experimental treatment (FIE) and the experimental design methodology, for example, subject assignment (random vs. nonrandom, parallel group design), instructor effects (the same teacher vs. different teachers for experimental and control groups), temporal effects (concurrent experimental and control classes vs. classes taught in different semesters or years), test author bias (standardized vs. teacher-developed tests), test scoring bias (objective test vs.  53  subjective test) and statistical control (statistical techniques to isolate or subtract variance i n the dependent variable attributable to variables that are not the subject o f study, for example, using an Analysis o f Covariance). Study settings and publication features w i l l also be recorded. These include course content (variations in F I E lessons), grade level, ability level o f population, year o f report and source o f report (unpublished, dissertation, published), author and country. A s suggested above, the quality o f the study can also be recorded. Data collection w i l l be completed when all studies are in. However, at this time it is likely that, in addition to data selection from the above, the following variables w i l l be recorded because they refer to Feuerstein's goals and to the theoretical and empirical evaluations o f the program: ethnicity, gender differences, socioeconomic status, age, population (for example, deaf, gifted, the educable mentally handicapped, learning disabled, behaviour disordered, and emotionally handicapped), length o f instruction, and the focus o f measurement (for example, verbal and nonverbal ability, Full Scale IQ, general knowledge, spatial perception, academic achievement, behaviour, self-esteem and concept, and motivation to learn) and the teacher variable (for example, teacher training and skill level).  Statistical Analysis Analyzing the data statistically usually includes combining the data to arrive at a summary estimate o f the effect size, a measure o f its variance and its 95% confidence interval, and a test for homogeneity o f effect size. It may include examination o f the reasons for heterogeneity (Petitti, 1994).  Choosing Estimates of Effect. Although it is rare for studies to present only one estimate o f effect for a meta-analysis, Petitti (1994), like K u l i k and K u l i k (1989), believes that one and only one estimate o f effect from each eligible study should be used i n the calculation o f the summary estimate o f effect, because  54  using more than one would inappropriately weight studies with many estimates o f effect size. However, several analyses w i l l be performed as, for example, only effects o f reading ability w i l l be extracted, then only effects o f mathematics ability, and so on. The analyses w i l l initially be chosen to address the research questions advanced i n Chapter Two.  Each variable to be estimated  w i l l be analyzed for homogeneity. If the test for homogeneity is rejected, a decision w i l l be made to either set the variable aside or to use the random effects model to analyze the data.  Calculations for the Unbiased Estimator of Effect Size for Each Study. The most familiar measure o f effect size, and the one chosen for this meta-analysis; is the standardized mean difference between outcome scores o f experimental and control groups divided by the standard deviation (Cohen, 1988; K u l i k & K u l i k , 1989; Petitti, 1994; Wolf, 1986). There are other methods that test the significance o f combined results, with Fisher's (1932) procedure being perhaps the one most widely used. However, while Fisher's procedure produces an over-all p-value, it does not provide an estimate o f the magnitude o f the effects being considered, which was a major focus o f this study (Hedges & O l k i n , 1985). In this meta-analysis, procedures from Hedges and O l k i n (1985, pp. 81-91, 109-113, 122-128) w i l l be used when means, standard deviations and the number o f subjects are reported. The fixed-effects model w i l l first be used because it is appropriate to answer the question o f whether or not the treatment has caused an effect in the studies that have been done. The alpha level for the test for homogeneity w i l l be set at 0.10. If the fixed-effects model is not successful, the data w i l l be re-analyzed using the random-effects model. Procedures from Hedges and O l k i n (1985, pp. 191-199) w i l l be used. If means, standard deviations or sample sizes are not given, effect sizes w i l l be calculated following Cooper (1998, p. 129) and W o l f (1986, p. 35) when /- and F-statistics are reported.  55  In most cases, the effect size can be positive or negative. A minus sign with a significant test indicates that the experimental group had results lower than those o f the control group. However, F-test results w i l l have a positive sign, and Mests can be calculated as Treatment minus Control or Control minus Treatment. Information from each study w i l l be used to decide whether a positive or a negative sign w i l l be used for the effect size. The unbiased estimator o f effect size, J(m), w i l l be used to reduce bias due to sample size differences. When descriptive statistics are not reported and variances cannot be calculated, it w i l l not be possible to combine the studies, calculate the test o f homogeneity, calculate the z variable and />-value o f d, or create a confidence interval for d.  Interpretation of Effect Sizes. According to Hedges and O l k i n (1985), because the effect size, d, is the standardized score (z score) o f the experimental group mean i n the control group distribution, the quantity, d represents the proportion o f control group scores that are less than the average score in the experimental group. Thus an effect size o f d = 0.5 implies that the score o f the average individual in the experimental group exceeds that o f 69% o f the individuals in the control group. Similarly an effect size o f -0.5 implies that the score o f the average individual i n the experimental group exceeds that o f only 3 1 % o f the individuals in the control group (pp. 76,77). Similarly, i f the d value is 0.32, and we consult a Normal Distribution (z) Table, we find that the area under the normal curve is 0.6255. This means that the average person receiving the treatment would have a score, say on a reading test, greater than 62.55% o f the individuals i n the control group. The treatment could be expected to move the typical person from the 5 0 to the th  62.55 percentile on a measure o f reading (Wolf, 1986). Another way o f stating this would be to th  56  say the 62.55% o f the experimental group, would exceed the mean o f the control group i f both were normally distributed (Schmidt, 1992). Cohen (1988) discussed the dilemma o f interpreting effect sizes that cannot be calculated with means and standard deviations. H e considered the effect size d = 0.20 to be small. In such cases, two normally distributed populations o f equal size and variability would have only 14.7% o f their combined area not overlapped. This effect size represents, for example, the approximate difference in mean height between 15- and 16-year-old girls. Cohen wrote that i n new areas o f research, effects sizes would likely be small or zero. This was because the phenomena being studied were typically not under good experimental or measurement control, or both. Cohen considered d = 0.50 to be medium-sized. In these cases, 33.0% o f the two populations would not be overlapped. He wrote that such an effect size would be large enough to be visible to the naked eye. In this case it represents the approximate difference in height between 14- and 18-year-old girls. Cohen considered d = 0.80 to be large with the two populations so separated that 47.4%, or almost half, o f their areas are not overlapped. He wrote that the differences are easily perceptible, as with the mean difference in height between 13- and 18-year-old girls. Chapter Four w i l l report on the methods that were used for this analysis.  57  Chapter Four  Methodology Studies were identified by a search o f personal reference files and by computerized searches o f the Educational Resources Information Centre (ERIC), Psychological Abstracts, Dissertation Abstracts and the Internet for published and unpublished reports, dissertations, and other fugitive literature. Search terms were "Instrumental Enrichment" and "Feuerstein's Instrumental Enrichment." Search dates were from 1966, the beginning date for the on-line indexes, until 1999, after which time I began to perform the calculations for the meta-analysis. The title and abstract o f studies identified in the computerized search were scanned to exclude any that were inappropriate based on criteria outlined in the following section. The full text o f the remaining articles was retrieved, and each paper was read to determine whether it contained information about an experimental or observational study o f FIE. The reference lists o f these articles were reviewed to identify citations to other studies o f FIE, and works that were not identified in the computerized literature search were retrieved and reviewed for presence o f relevant information. Pertinent data published in languages other than English were identified and individuals connected to research into F I E were personally contacted v i a e-mail. In this manner 36 studies were accepted and 20 rejected. O f the accepted studies, all were quasi-experimental, 11 were published, 15 were Doctoral dissertations, three were Master's theses and seven were unpublished reports. Twenty-two were from the United States, seven from Canada, three from Israel, two from Britain, one from South Africa and one from France. Publication dates ranged from 1979 to 1996. Inclusion-exclusion criteria are given below.  Criteria Used to Include or Exclude Specific Studies Studies were chosen only i f the all three o f the following criteria were met: 1) Feuerstein's Instrumental Enrichment (FIE) was the intervention that was used,  58  2) their experimental designs included post-test measures for each subject, and 3) both treatment and control groups were used. This last condition was necessary because this analysis was designed to compare scores of treatment and control subjects on post-treatment measures only. In addition, i f a study included treatments in addition to FIE, and i f there was a control group, post-test scores for only the F I E group were compared to post-test scores for the control group. In cases where the same study was reported in multiple publications, only one set o f results, the most comprehensive set, was chosen. Subjects ranged in age from 9 to 20, although one study contained subjects who were Technical School students training to be masons, and some could have been older than 20. The Summary o f Studies Used, Measures and Effect Sizes (Appendix A ) contains a summary o f the 36 studies that were included, details regarding their design, the dependent measures used and the effect sizes calculated from each study's data. Dependent measures were included i f they evaluated subjects' performance on standardized measures. This was because most studies used standardized measures, and many non-standardized measures, such as interviews and attendance figures, were impossible to rigorously quantify. In addition, informal, locally developed measures such as teachers' and parents' ratings may not have contained objective opinions. Standardized measures have been validated on groups o f individuals, and given that there was such a large amount o f variance among the studies i n this analysis, the use o f standardized tests makes the results more reliable . The standardized 1  measures were o f full scale, verbal and nonverbal reasoning ability, visual perception and visual-motor ability, academic achievement, self-confidence, self-esteem, self-efficacy, and  A caveat regarding the interpretation of results from standardized tests with the FIE population will be presented in Chapter Six.  59  self-concept, and attitudes and behaviour. It was possible to collect enough data from these measures to calculate 360 effect sizes. The meta-analysis did not include all dependent measures from included studies, and detailed reasons for their omission are given in Appendix A . Briefly, within the included studies, four measures were not included because they did not provide data on how F I E impacted students' performance on the standardized measures described above. Thirty measures were not included because they were informal measures that did not report reliability, validity or standardized scores. Thirty measures were not included because the data were reported i n ways that could not be used i n the meta-analysis. Examples o f this include the following: 1. missing data on the number o f subjects; 2. basic descriptive information such as means or standard deviations were not reported; 3. pre-post results for the treatment groups were reported, but the post-test results were not compared to a control group; 4. the outcome measures were administered to only the treatment group, but not to the control group; 5. a single-subject research design was used (this is a quasi-experimental research design that involves assessing change i n a dependent variable on a single research subjects); 6. non-parametric statistical procedures were used. The Mann-Whitney U non-parametric statistical procedure was not included because data were reported i n terms o f ranks and it was not possible to calculate effect sizes (Hedges & O l k i n , 1985). In addition, in four included studies, the authors reported only the results o f significant differences between the experimental and control groups but omitted any reference to nonsignificant differences.  60  M a n y studies reported several results from a variety o f different dependent measures. In these cases, only one effect size per study per domain was used. For example, one study might generate data used both in the analyses o f reading and the analyses o f self-esteem. Where possible, composite, or total, scores rather than individual subtest scores were used. The Summary o f Studies Not Used (Appendix B ) contains a summary o f 20 studies that were excluded from the meta-analyses and the reasons for their exclusion. A summary o f included and excluded studies is given i n Table 1, below. Nine other studies could not be included because they were Ph.D. dissertations or Master's theses, and the universities where they were housed either would not lend them or did not respond to the Inter Library L o a n request . 2  Table 1  Summary Of Included and Excluded Studies Included  Not Included  11 15 3 7 36  10 5 1 4 20  Published Doctoral Dissertations Master's Theses Unpublished Reports Totals:  Quality of Individual Studies. Regarding the quality o f the individual studies, all were quite similar i n that they were quasi-experimental studies. According to Campbell and Stanley (1963), "There are many natural social settings i n which the research person can introduce something like experimental design into his scheduling o f data collection procedures (e.g., the when and whom o f measurement), even though he lacks the full control over the scheduling o f experimental stimuli (the when and whom o f exposure and the ability to randomize exposures) which makes  The universities were Columbia University Teachers' College, Wayne State University, University of Pretoria (South Africa), Temple University, National-Louis University, Saint Xavier College and Northeastern University.  2  61  a true experiment possible. Collectively, such situations can be regarded as quasiexperimental designs" (p. 34). This situation arises whenever research is attempted in school settings where the researcher often has to take what is available i n the way o f schools, classes and subjects. Control over exposure and true randomization are, therefore, not available. Researchers do whatever they can, however, to control the research situations, and for the most part the studies in this analysis appeared to be o f adequate quality within the restrictions o f quasi-experimental design. Notes on study quality are given for each study i n Appendix A . The following is a compilation o f information that most study authors reported . Students were rank-ordered on IQ tests and alternately assigned to either treatment or control group. Students were pre-tested to ensure that there were no significant differences between groups on pretest measures. Students were randomly assigned from their regular classes to small core classes, or all students i n the participating grade level were randomly assigned. Comparisons were also made between matched pairs, or students were matched on the bases o f sex, age and level o f class placement or learning problem and then divided randomly into two groups. Results were often analyzed by analyses o f covariance where, for example, gender and school factors were covariates. Nested designs were also used where, for example, treatment was nested under classrooms. Testing personnel were often blind to the specific hypotheses o f the study. Treatment teachers were either experienced F I E teachers or had received training i n F I E instruction. In one large study teachers were randomly assigned to treatment or control conditions and the treatment teachers were then trained. However, i n spite o f the fact that the study authors attempted to compensate for the lack o f randomization, it is possible that the groups that were compared were not balanced with respect to potential confounding variation. In addition, because the studies i n this meta-  A minority of authors did not report enough information to enable an evaluation of study quality. This will be discussed in the limitation section of this dissertation. 3  62  analysis were not themselves randomly selected, we cannot assume with certainty that the results w i l l generalize to the larger population. The results, discussed in Chapter Six, should be interpreted with this in mind.  Statistical Methodology A s described in Chapter Three, the fixed-effects model was the first choice for analysis because it is appropriate to answer the question o f whether or not the treatment had caused a common effect across the studies that had been done. In order to use as many effect sizes from individual studies, calculations were divided into first, domain, such as reading, verbal ability, and so on, and then divided again into groups according to the length o f the treatment. Using this method, it was possible to use only 72 effect sizes from a total o f 360 to create 28 combined analyses. This means that 288 effect sizes were not combined. O f these 288, 228 did not meet the criterion for homogeneity in their particular domains (e.g., reading), two were from studies where effect sizes only were reported, and 58 were from studies where F- or ^-statistics only were reported. It can be seen from the above that more than half the effect sizes did not meet the criterion for homogeneity and could not be combined using the fixed effects model. Variables such as gender, ethnicity, socioeconomic status, age, intellectual ability, and physical, learning, emotional and behavioural disabilities were not adjusted for, likely contributing to heterogeneity within and across studies. The data were, therefore, analyzed again using the random effects model. In general length o f treatment was not considered. However, i f a researcher used the same subjects for treatment and testing for 1 year, again after 2 years, and finally testing without treatment after 3 or 4 years, the different years were not combined separately into the larger group because the analysis would not have independent groups o f subjects.  63  The File-Drawer Problem A s discussed in Chapter Three, the "file-drawer problem" is one i n which published research is biased in favour o f significant findings because nonsignificant findings are rarely published. This situation can lead to biased meta-analysis in favour o f significant, or positive, results. Care was taken i n this meta-analysis to choose unpublished as well as published studies, and o f the 36 studies included, 11 were published and 25 were not. In addition, a variation o f the Fail-Safe procedure, developed by Orwin (1983), was used to calculate how many additional studies with null results would be needed to reverse a conclusion that an observed effect size was significantly different from zero (see Hedges and O l k i n , 1985, p. 306). A sample o f three significant effect sizes was chosen from the results presented i n Appendix C . Verbal Ability had a large effect size, 1.4050, and following Orwin, 78 additional studies with null results would have to be added to reduce that effect size to the negligible size o f 0.10. Visual Perception had an effect size o f 0.4237, slightly less than medium. Thirteen additional null results would have to be added to reduce it to 0.10. General Academic Achievement had a small effect size, 0.2571, and 11 additional null results would be required to reduce it to 0.10. It is unlikely, even for the small significant effect sizes, that the required number o f studies with null results exists for these examples, and therefore, it is unlikely that the results o f this meta-analysis were influenced by the file-drawer problem. Nevertheless, as w i l l be discussed in Chapter Six, the choice o f the random-effects model to analyze data can increase the chances o f publication bias, and the results should be considered with this in mind.  The Hawthorne Effect Results o f the meta-analysis were examined to see i f significant positive results were due to "The Hawthorne Effect." The Hawthorne Effect is one in which treatment groups  64  perform better than control groups simply because they have been given more, or more stimulating, attention, and they have responded to that. In this meta-analysis 28 studies had regular instruction for the control groups and 8 had novel activities. A n examination o f the combined studies that produced significant effect sizes revealed no direct connection between the type o f control group and whether it had an influence on significance. For example, a breakdown i n the F u l l Scale Ability table i n Appendix D reveals that o f the eight studies involved in that case, one provided a novel situation to the control group. That study produced a significant effect size for the treatment group when the logic o f the Hawthorne Effect suggests that it should not necessarily have done so. O f the remaining seven studies i n that table, none had novel control groups, and three o f them produced significant effect sizes while four did not. There is no clear pattern showing significant results when studies had control groups that did not receive novel treatment. This same mixed pattern appeared i n the other tables i n the summary. Results o f the statistical calculations are presented i n Chapter Five and discussed in Chapter S i x .  65  Chapter Five  Results The research question, (number 5) as stated on pages 32 and 33 in Chapter T w o is, H o w far does Instrumental Enrichment transfer? That is, specifically: a) A r e there significant effect sizes for cognitive tasks such as full scale reasoning ability, and verbal and nonverbal reasoning ability for students who have participated i n Instrumental Enrichment programs? b) A r e there significant effects for measures o f academic achievement and in the affective domain on measures o f self-esteem, self-confidence, self-concept, motivation to learn, behaviour and attitudes? c) A r e significant effect sizes restricted to tasks similar to those i n the F I E curriculum, for example, visual-spatial tasks, visual-motor tasks and dynamic assessment with the Learning Potential Assessment Device?  Overall Effect Sizes In this analysis 360 individual effect sizes from 36 different studies were calculated. Fifty-eight were calculated from F- or /-tests, and two were from studies where effect sizes only were reported; these could not be incorporated into the subsequent calculations. Random Effects M o d e l analyses were performed on the remaining 300 individual effect sizes, and combined effect sizes were obtained for 22 categories. Detailed results o f the Random Effects analyses are presented i n Appendix C . O f the 34 combined analyses presented in that appendix, 12 had significant effect sizes and 22 did 4  not. Specific Tests used by individual authors within significant combined analyses are given in Appendix D . Their individual effect sizes are also given i n that appendix, along with a  4  These can be a combination of studies or a combination of tests or subjects from a single study.  66  statement as to whether or not that effect size was significant. Nonsignificant combined results are presented in a similar manner in Appendix E . Effect sizes from F- and Mests are given in Appendix F .  Near and Far Transfer Significant results from the combined analyses (see Appendix C ) are grouped below according to near and far transfer. Although there are several significant effect sizes from a single administration o f one test, effect sizes from combined analyses only w i l l be reported, as results from a single administration may not provide reliable information. When examining these results it should be remembered that, as discussed i n Chapter Three, Cohen (1988) considered an effect size o f 0.20 to be small, that o f 0.50 to be o f medium size, and that o f 0.80 to be large. The results are summarized i n Table 2 at the end o f this chapter, and implications o f the findings w i l l be discussed in more detail i n Chapters Six and Seven. The first group o f significant results consists o f those from measures considered to be similar (or near) to the activities and tasks presented in the F I E curriculum. They are measures grouped into the categories o f Visual and Visual-Motor Ability which are a major focus o f the F I E curriculum. There were significant results for Visual Perception after five months to two years o f treatment (J=0.42, p<.05), Visual-Motor Ability after one-half a year o f treatment (J=0.71, p<.01), Visual-Motor Ability 1 year after a year-long treatment (d=1.68, p<.01). Significant results on near transfer effects are also drawn from measures from the cognitive domain, that is, from tests combined into the categories o f F u l l Scale A b i l i t y , Verbal A b i l i t y and Nonverbal Ability. The results show that there were significant results for F u l l Scale A b i l i t y after 1 and 2 years o f treatment (J=0.58, p<.01), Verbal A b i l i t y after 4 weeks to two years o f treatment (d=\A\,  p<.01), and Non-Verbal A b i l i t y after 3 months to 2 years o f  treatment (rf=0.24, p<.01).  67  The results from measures o f academic achievement are considered to be i n the middle o f the continuum between near and far transfer because, although cognitive skills such as memory and reasoning are involved, the contents o f the achievement tests were not directly taught within the F I E curriculum. These results showed significant effect sizes for General Academic Achievement after 4 months to 2 years o f treatment (d=0.26, p<.05), and M a t h Achievement after 1 and 2 years o f treatment (d=0.29, p<.05). The next results are from Jensen's (n.d.) study and are difficult to interpret because, with the significant results, he did not report information on individual tests. H e reported only that near and far transfer tasks were selected from the French K i t and frOm "up-coming and more advanced parts o f the IE curriculum" (pp. 14,15). "Near Transfer" was significant after one and one-half years o f treatment (d-\.6l, p<.01) and "Far Transfer" after one and one-half years (d=l.08, p<.01). The last group o f significant results come from the affective domain, that is, from measures o f personal behaviour, feelings and thoughts. They are far transfer measures because the contents o f the tests were not similar to those in the F I E materials. There was a significant effect for Locus o f Control-Intellectual Achievement after four and one-half months to one year o f treatment (d-0.33, p<.05). There was also a significant negative finding for General Motivation, Attitudes and Motivation for Learning 1 year after the end o f 2 years o f treatment (d= -0.48, p<.05). Negative findings indicate that the control groups performed better than the treatment groups. To summarize positive combined results, effect sizes for near-transfer measures o f visual perception ranged from less than medium to large. They were mixed for near-transfer measures o f cognitive ability, ranging from large for verbal ability, to medium for full-scale ability, and small for non-verbal ability. There were small to less than medium effect sizes for  68  mid-transfer measures o f general achievement and math achievement. The intellectual achievement locus o f control effect size, a measure o f far transfer, was less than medium. Results for the following were not significantly different from zero: F u l l Scale A b i l i t y Perry and Samuels with 2 years o f treatment; Negative Visual Perception; Learning Potential Assessment Device ( L P A D ) ; Reading Academic Achievement; M a t h Academic Achievement with 2 years o f treatment and 1 year after treatment ended; Self-Esteem and Confidence i n Personal Success; Academic Self-Confidence; Self-Confidence o f Physical Competence and Appearance; Social Acceptance, Skills and Popularity; Self-Concept; General Motivation and Attitudes and Motivation for Learning - 4 weeks to 2 years and 2 years; Behaviour; and Negative (or poor) Behaviour. O f the nonsignificant results listed above, the majority are from the affective domain. In addition, one effect size for General Motivation, Attitudes and Motivation for Learning 1 year after the end o f treatment was significantly negative. Both significant and nonsignificant results reveal important information regarding the performance o f F I E and w i l l be discussed i n Chapter Six.  69  CD CJ  CS  c  e  o  a.  3  5  a  cj  il  C  o  IB  co  _o  SB  60 C  I  C3 CD  CJ  >  15 a.§in Sc a sQ. '5O iS T3 i in  c o=  :  O  C O  CJ  2^  cs  1 .5E  •o  7 e  lcs  <  t/3  o  &  CJ ^ n-. -—> T j CJ  8"S  U  J  C3  CD "cS  CD  <c  (/I  CD T3 3 -*-»  c  CJ  o  I  CN  CN  C CS  a. T3  £ o  —  T3  O  a.  ^  u  _CS  <u •*-» CD  D  c  cs  c  CS  •3O  O 22 <D  -CS 3  c  CO  5  .2 >  2  oo O b  CN  £3 •CSl i  b  03  o  5;  C/5  CN CJ  o  cb  g  cs  ca u £>CN  'S  U,  o  CN  >>  e o  . ^  , aftei  CN  |  a.  <u o CJ  U cO X)S  3  •J3 D. tu  "  >> cS — i_ O o  a.  2  o co  cs cs 3 Ul  |  S  S  O N !  <N 1  J2  3 «  cS  £  3  a  >  CS CS ° -2 'c —3 h U  b S ra U  CS * J  » CL,  > > > b n- >  T3 CJ •a  s  _3  "3 .3 O  a  «  E CN  cd H  'a  .2f 53  o  a  a  c  £'3  CJ  c  OX)  CS  c o  3D be  CS CJ  J3 C/5  O  8 j 70  Chapter Six  Discussion The research question, stated in Chapter Two, is as follows: H o w far does Instrumental Enrichment transfer? That is, specifically: a) A r e there significant effect sizes for cognitive tasks such as full scale reasoning ability, and verbal and nonverbal reasoning ability for students who have participated i n Instrumental Enrichment programs? b) A r e there significant effects for measures o f academic achievement and in the affective domain on measures o f self-esteem, self-confidence, self-concept, motivation to learn, behaviour and attitudes? c) A r e significant effect sizes restricted to tasks similar to those i n the F I E curriculum, for example, visual-spatial tasks, visual-motor tasks and dynamic assessment with the Learning Potential Assessment Device? The F I E curriculum consists o f lessons with many abstract visual stimuli such as dots, diagrams, drawings and patterns o f shapes. The lessons also involve having students analyze relationships among these stimuli and describe and explain these relationships. Therefore, for the purposes o f this study, the term "near transfer" w i l l refer to assessments or tasks requiring similar activities, that is, the examination and rational discussion o f abstract visual stimuli and of ideas. "Far transfer" w i l l refer to more dissimilar activities such as behaviour, attitudes, self-confidence and self-esteem. Academic achievement is in the middle o f the continuum. This discussion w i l l be confined to effect sizes from combined analyses only, and Jensen's (n.d.) "near-" and "far-transfer" results w i l l not be discussed since, without knowing what the exact tasks were, it is not possible to know how closely they resembled those i n the F I E curriculum. In general, results o f this meta-analysis support the observation that  71  successful performance on the many measuring instruments used by researchers follow a nearto far-transfer pattern. There are more significant results for near-transfer measures than for mid- and far-transfer measures, and for the most part, they are larger. Before discussing the results, it might be argued that subjects i n the studies improved, not because o f F I E , but simply because they were older when the treatment ended and the apparent gains were due primarily to development. However, the developmental factor was controlled with the use o f same-aged groups. The following caveats should be considered while reading the discussion o f results in this chapter and the implications o f those results i n Chapter Seven. A s was discussed in Chapter Four, because the studies used for this meta-analysis were quasi-experimental and not themselves randomly selected we cannot confidently assume that the results w i l l generalize to the larger population. Regarding the decision to analyze standardized tests, these measures are normed on a large population, and it is likely that the majority o f that population consists o f individuals who are from average-level living conditions. Many o f the IE groups consist o f individuals from disadvantaged circumstances, and the standardized measures may not be valid and reliable with this population. For example, regarding the l o w results i n the affective domain, there may have been positive changes, but the tests measuring that domain may not have been appropriate and did not correctly measure these changes. Therefore, caution needs to be used in interpreting the results. Regarding the choice o f the random effects model to analyze the data, it has been reported that up to 50% o f meta-analyses can suffer from publication bias (Sutton, Duval, Tweedie, Abrams & Jones, 2000). The random effects model, because it puts heavier weight on the smaller studies, is more susceptible to publication bias than the fixed model. This is  72  because publication bias mainly implies that small sample null studies are not likely to be published and are, therefore, missing from a random effects meta-analysis. It is, therefore, possible that my estimates may be somewhat high because some small sample null studies are not i n my analysis.  Discussion of Results Near Transfer. The highest overall collection o f significant effect sizes is for Visual Perception (d=0.42) , Visual-Motor Ability (d=0.71), and Visual-Motor A b i l i t y 1 year after treatment 7  ended (d=l .68). These results reflect the emphasis on visual-spatial material i n the F I E curriculum. In the case o f the follow-up result, visual material skills appear to have been maintained. However, this effect size was based on the results o f only two tests and it may not have been as large i f more tests could have been included. The next group consists o f Non-Verbal Ability (d=0.24), F u l l Scale A b i l i t y with 1 and 2 years Of treatment (d=0.5S), and Verbal A b i l i t y (d=l .41). Although the F I E curriculum presents material visually, it also requires verbal analysis and discourse, and the high verbal ability effect size likely reflects that. Tests included in the non-verbal ability calculations were not the same as those for the visual perception calculations described above (see Appendix D ) . The tests required reasoning in addition to visual skills. Nevertheless, it is somewhat surprising that the verbal ability effect size is so much larger than that for non-verbal ability. Near transfer results were not uniformly significantly positive. Calculations for the following did not produce significant effect sizes at p<=.05: F u l l Scale A b i l i t y with 2 years o f treatment, Negative Visual Perception (these were error scores), and the Learning Potential Assessment Device (LP A D ) with 1 year o f treatment, and 1 year after treatment. The L P A D  7  See Chapter Five  73  results were close to being significant (d=0.50,p=.06) in the analysis that combined several studies, but the effect is o f only medium size. The result is puzzling given that Feuerstein and other authors who supported his theories developed the L P A D , and the materials are similar to those i n F I E . The result suggests the need for further study with the L P A D to evaluate F I E .  Mid-Transfer. Results for measures o f academic achievement were mixed. Calculations for the following produced significant effect sizes: General Academic Achievement (d=0.26) and Academic Achievement, Math with 1 and 2 years o f treatment (d=0.29. Although significant, they are fairly small. Calculations for the following were nonsignificant: Academic Achievement, Reading, with 1 and 2 years o f treatment, with 2 years o f treatment, and 2 year after treatment ended; Academic Achievement, Math, with 2 years o f treatment and 1 year after treatment ended. The last two results, for Math, had negative effect sizes that were close to being statistically significant. These results are consistent with the near- to far-transfer pattern where subjects perform better on tasks similar to the content o f the F I E curriculum and worse as the tasks become more dissimilar. In Chapter One various theories o f critical thinking were considered, and it was pointed out that there was debate about whether critical thinking is subject specific or generalizable across subject areas (see pp. 5-6). Feuerstein belongs to a group that believes there are general principles o f critical thinking and that these should be taught separately from standard subject areas. However, results from this meta-analysis suggest that these teachings do not generalize well to other subject areas. This situation w i l l be discussed further i n Chapter Seven.  74  Far Transfer. Calculations produced a significant positive result for Locus o f Control - Intellectual Achievement (d=0.33). This is a fairly small effect size; however, it suggests that the students felt they were responsible for their intellectual achievement. There was one other significant effect size, but it was negative, indicating that the control group had higher scores than the treatment group. It was for General Motivation, Attitudes and Motivation for Learning 1 year after treatment ended (d= -0.48). The result was from only one study, however, and may be a reflection o f the particular population o f students being tested. Effect sizes were nonsignificant for General Motivation, Attitudes and Motivation for Learning i n the combined group with 4 weeks to 2 years, and with 2 years, o f treatment. Calculations for the following also resulted in nonsignificant results: Self-Esteem, Confidence in Personal Success, Academic Self-Confidence, Self-Confidence o f Physical Competence and Appearance, Social Acceptance, Skills and Popularity, Self-Concept, Behaviour with one year and 2 years o f treatment and 1 year after treatment, and Negative (that is, poor) Behaviour with 1 and 2 years o f treatment and 1 year after treatment. The above results are important i n that they are consistent with the anecdotal reviews of researchers such as Sternberg (1984) and Sternberg and Bhana (1986). A s reported i n Chapter T w o , Sternberg and Bhana believe that F I E trains primarily those abilities that IQ tests tap rather than a broader spectrum o f abilities that go beyond intelligence as the tests measure itSimilarly, Savell et al., (1986) noted that F I E effects most commonly reported were those on certain standard nonverbal measures o f intelligence that are largely measures o f skill in processing figural and spatial information. They commented that effects on other types o f measures, such as self-esteem, impulsivity, classroom behaviour, academic achievement and  75  course content were absent, inconsistent or difficult to interpret. The results in this metaanalysis support the conclusions o f these authors. Regarding the contention o f Frisby and Braden (1992) that social intelligence is highly modifiable (see Chapter Two, p. 28), these results suggest that social intelligence needs to be taught i n a more direct manner in IE so that it becomes a near-transfer variable that might be modified more than it appears to have been.  Subgoals of Instrumental Enrichment. Although they vary, the significant positive results i n this meta-analysis appear to support the first two subgoals o f F I E as described at the beginning o f Chapter T w o . The first subgoal is the correction o f deficient cognitive functions that can be in the input phase o f information processing where necessary information is gathered, the elaboration phase where the gathered information is used, or the output phase where the solution to a problem is expressed. The second subgoal is the enrichment o f cognitive repertoire, including the vocabulary, cognitive operations, strategies, concepts and relationships necessary to complete F I E tasks. The lack o f success on measures in the affective domain is interesting i n view o f the last four subgoals o f the F I E model. They have to do with the affective domain, that is, production o f intrinsic motivation, the production o f reflective, insightful thinking processes and the change i n students' self-perception from being passive recipients o f information to that o f being creative originators o f new information. The poor results, as reflected i n this metaanalysis, could be because, as Sternberg and Bhana (1986) state, those attributes are not built into the instruction contained in the program. If they were, their measures would fall into the near-transfer category and positive results might increase.  76  Chapter Seven w i l l provide further discussion as well as suggestions for future research related to the questions and results i n this meta-analysis.  77  Chapter Seven  Implications for Practice and Research This meta-analysis and its results lead to implications for practice and implications for research. Regarding educational practice, the results suggest educators can be reasonably confident that the IE curriculum can create positive changes in students that produce, on standardized measures: •  large effects for visual-motor ability and verbal ability.  •  medium effects for visual perception and full-scale ability, and  •  small effects for non-verbal ability, general academic achievement, math (perhaps) and for a feeling o f being responsible for their intellectual achievement. I believe, however, that the most important implications arise, not from the significant  results, but from the nonsignificant ones. These w i l l be discussed i n more detail below.  Implications for Practice Affective Domain Instruction. The lack o f positive effects in the affective domain suggests that there is a need to include direct affective domain instruction into IE. In their book on experiential learning and change, Walter and Marks (1981) discuss a rational restructuring process built upon models o f cognitive psychology, not cognitive ability. Describing one aspect o f the change process they say: The term re-cognition is introduced here and used to categorize approaches and mechanisms for changing the way an individual thinks. For example, internalized problem solving sequences, altered mental verbalizations and imagery, altered constructs for organizing, and even hypothesis formulation are included i n this discussion, (p. 10) 78  This quotation sounds very similar to those o f Feuerstein and his colleagues, yet it refers to the affective, not the cognitive, domain. F I E encourages students to ask questions, reason inductively and deductively, give clear solutions to problems i n an empathetic manner (that is, putting themselves "into the shoes" o f listeners to be sure the answer w i l l be understood), justify their opinions, restrain impulsivity and examine relationships in cognitive situations. However, it does not directly teach students how to work with emotions or attitudes that may arise as a result o f the "cognitive" instruction. There are cognitive-behavioural methods in psychology that can be used. For example, Stein and Book (2000) provide exercises to assess and improve performance in the Intrapersonal, Interpersonal, Adaptability, Stress Management and General M o o d realms. B y adding a direct instructional component for the affective domain, F I E would be introducing a way to address the far-transfer problem. When researchers measure students' performance in that domain, their measurements would then be o f near-, not far-transfer tasks, and they could provide more positive results.  Counselling. The lack o f positive results in the affective domain also suggests that there is a need to attach a counselling component to IE instruction. In his text on counselling skills, Egan (1998) cites several obstacles to effective client self-exploration. He states that, "Digging into one's inadequacies always leads to a certain amount o f disequilibrium, disorganization, and crisis. But growth takes place at crisis points" (p. 139). Although the F I E curriculum does not "dig into one's personal inadequacies," as stated above, it does encourage questioning, reasoning, and giving and justifying opinions. It may be that the l o w affective domain effect sizes are a result, not o f a defect i n F I E , but that it is succeeding to a point where it is bringing students to that personally felt crisis point where long-lasting change can happen - or be avoided. In  79  addition to direct affective-domain instruction, personal counselling should be offered outside of the classroom during the F I E program and for some while after it ends. Another argument in favour o f offering counselling to students is that l o w behaviour, motivation and attitude scores could also come from students questioning authority as a result o f FIE. Because F I E is customarily taught during early and late adolescent years, it coincides with a time when adolescents are struggling with issues o f identity. According to M a r c i a (1966,1980) and Marcia, Waterman, Matteson, Archer, and Orlofsky (1993) young people go through a period o f crisis and decision-making, and depending on the decisions, may, or may not, achieve full autonomous identity. F I E may enhance, hinder or confuse the process.  Academic Instruction and IE. Attempts should be made to insert IE instruction into specific academic subject areas, and the IE instruction should relate to those subjects. In Chapter Six results o f mid-transfer calculations were discussed. General Academic Achievement and one instance o f Math achievement had significant, although small, effect sizes. Other measures o f M a t h and Reading did not have significant effect sizes, indicating that instruction in IE did not generalize, or transfer, well to academic instruction. A s discussed i n Chapter One, although Feuerstein belongs to a group that believes the general principles o f critical thinking should be taught separately from other subject areas, other researchers disagree. Ennis (1985b) and Sternberg (1987) recommend the use o f a mixed model o f providing a separate critical thinking course along with the insertion o f critical thinking approaches into specific subjects to reinforce all skills taught. This approach could be used where the IE curriculum is taught separately while at the same time, its methods are used i n subjects such as math and reading. In this situation, math and reading assessments might become more near-transfer tasks and students could have higher performance scores.  80  Implications for Research The LP AD More studies should be carried out using the L P A D to measure success i n Instrumental Enrichment. In this meta-analysis, L P A D results were not statistically significant, and as discussed i n Chapter Six, this is surprising since Feuerstein and his colleagues designed it, and many subtests have materials that are similar to those i n IE. The L P A D is a "dynamic" test with three steps. A subtest is first administered to measure initial levels o f performance (a baseline). Second, the assessor gives "training i n the form o f the teaching o f principles o f thought and problem solving as well as modification o f specific cognitive process that may be required in the learning itself (mediated learning experience)" (Feuerstein et al., n.d., p. 2.2). The third step is the administration o f a similar form o f the first subtest. The expectation is , that there w i l l be an improvement i n performance. Scores are not compared to those o f a norming group but are considered to be a measure o f the subject's potential level o f cognitive functioning. Researchers such as Tzuriel and Alfassi (1994) have recommended that i n further Instrumental Enrichment research, a differentiation should be made between performance on static instruments such as IQ tests and on dynamic measures o f performance. If such research includes the use o f the L P A D the results should increase our knowledge o f not only Instrumental Enrichment effectiveness but o f whether or not it is most effectively measured with a static or a dynamic measuring instrument. Six tentative research questions were proposed in Chapter T w o . They arose from the literature review and represent concerns voiced by researchers and reviewers. This metaanalysis explored one: " H o w far does Instrumental Enrichment transfer?" Future research should explore the other five questions to fully evaluate the effectiveness o f IE.  81  Teacher Effectiveness and Bridging. The first two research questions listed in Chapter Two (see p. 32) have to do with teachers. The first is: "What is the optimal amount o f teacher training required i n order to best teach Instrumental Enrichment?" The second is: " H o w skilled were teachers with a mediational teaching style and with bridging, and did this make a difference to student outcomes?" F e w studies have been conducted on the effectiveness o f IE teachers, and the following discussion highlights the need for them. In his comprehensive evaluation o f F I E , Blagg (1991) referred to Annett (1989) who drew a distinction between "transferable" skills and "transfer" skills. They describe transferable skills in a manner similar to those o f other authors i n Chapter One when they discussed l o w road transfer that usually occurs in near transfer situations (see p. 7). They are commonly occurring skills and routines required i n many different contexts. The selection and organization processes involved in using these skills are under the control o f transfer skills that Annett claims refer to higher level, or executive, metacognitive processes. A s discussed i n Chapter One, these are analogous to the high road transfer often occurring i n far transfer situations. Annett believes it is one thing to demonstrate the acquisition o f transferable skills in particular contexts but another to have the transfer skills to know when to select and use these lower order processes i n novel situations. Blagg (1991) reported that in his study teachers and students complained about too much repetition with a lack o f novelty and stimulus. Once students had become aware o f implicit rules and learned how to approach the earlier tasks in each curriculum instrument, the later task pages needed little problem analyses or debate. He wrote that in some cases this created boredom, but more importantly, "what remained was the practice o f particular procedures (such as systematic search, counting, labelling, and eliminating) tied to artificial  82  exercises remote from everyday experience. Thus, the F I E program was often reduced to training in transferable skills rather than training for transfer" (p. 147). Although the teacher's manual did include many examples o f transfer or bridging applications, insufficient challenge, too much repetition, and artificiality in the tasks led to pupils and the teacher not always taking bridging seriously.  Further Analysis of Data in this Meta-Analysis. Question Three in Chapter Two asks: "What is the optimal length o f time that should be spent teaching Instrumental Enrichment for best results?" Question Four asks: " H o w many instruments are needed to produce optimal results?" The last question, six, asks: "What types o f populations benefited from IE instruction? Populations that have been studied include those related to gender, ethnicity, socioeconomic status, age, physical handicaps, intellectual ability, learning disabilities, and emotional and behavioural handicaps." It was not possible to analyze these factors in this meta-analysis because o f the complicated nature o f the data, time constraints and the choice o f methodology. However, Appendix A contains all the information necessary to analyze length o f time, number o f instruments and population data. There are also quantitative procedures other than the one chosen in this meta-analysis for analyzing results from existing studies, and any o f the six research questions could be analyzed using these procedures.  Brain Research. Studies into the effectiveness o f cognitive instruction, particularly that focusing on near- and far-transfer effects, may increase the body o f knowledge related to the study o f neuroplasticity. Neuroplasticity refers to the process i n which the brain changes physiologically as a result o f experience, and several researchers are investigating that process in relation to education (for example, Brandt, 1999; Wolfe and Brandt, 1998). It may be  83  possible to demonstrate a physical parallel to near- and far-transfer effects on tasks by using brain-imaging technologies. For example, Bigler, Lajiness-O'Neill, and Howes (1998) published P E T (positron emission tomography) photographs o f the brain being activated in four separate areas while hearing words, seeing words, reading words and generating verbs. If studies can discover which parts o f the brain are activated during specific cognitive or academic tasks, then instruction can be directed more efficiently by using near-transfer teaching techniques to activate particular areas o f the brain.  Limitations of this Study Limitations of, and problems with, this meta-analysis suggest ways in which metaanalytic procedures could be improved. Effect sizes were not reported i n most studies, and it was difficult to calculate them because o f the way results were reported. Specifically, statistical descriptions often did not report means, standard deviations or sample sizes. In addition, some authors did not report non-significant statistical results, making it impossible to calculate effect sizes. Individual studies should always report effect sizes and confidence intervals around them as well as their /rvalue. The method by which they were calculated and descriptive statistics should also be reported so their accuracy can be verified. The Fifth Edition o f the Publication Manual of the American Psychological Association (2001) recommends these procedures, as does the Task Force on Statistical Inference o f the American Psychological Association (Wilkinson and Task Force, 1999) In some studies information was not reported on pupil or Instrumental Enrichment selection procedures or on the teachers' level o f training. It was difficult, therefore, to judge the quality o f the studies. Study quality issues should always be reported. Regarding this meta-analysis, I found it difficult to get some studies. For example, several universities would not lend theses or dissertations and it would have been prohibitively  84  expensive to purchase them. Some universities also did not respond to Inter Library Loan requests (see Chapter Four, p. 60). Interested researchers could enlarge and re-calculate the meta-analysis with the addition o f information i n these missing studies. Finally, it is likely that the quality and implementation o f the programs used i n this meta-analysis varied considerably, coming from different countries and with different circumstances regarding schools, subjects and teachers. Therefore, it is probable that the program is not consistently the same from study to study. This is a curriculum implementation issue and is beyond the scope o f this study; it should be addressed i n future research.  85  References  American Psychological Association. (2001). Publication manual of the American Psychological Association ( 5 ed.). Washington, D C : Author. th  Annett, J . (1989). Training in transferable skills. Sheffield, England: The Training Agency. Arbitman-Smith, R., Haywood, H . C . , & Bransford, J.D. (1984). Assessing cognitive change. In P . Brooks, R . Sperber, & C . McCauley (Eds.). Learning and cognition in the mentally retarded, (pp. 433-471). Hillsdale, N J : Erlbaum. Asher, W . (1990). Educational psychology, research methodology, an meta-analysis. Educational Psychologist, 25, 143-158. Bachor, D . G . (1988). D o mentally handicapped adults transfer cognitive skills from the instrumental enrichment classroom to other situations or settings? Mental Retardation and Learning Disability Bulletin, 16, 14-28. Bailin, S. (1994). Critical thinking: Philosophical issues. In T. Husen & T . N . Postlethwaite (Eds). The international encyclopedia of education (2nd ed., V o l s . 1-11, pp. 12041208). N e w York: Pergamon Bigler, E . D . , Lajiness-O'Neill, R., & Howes, N - L . (1998). Technology i n the assessment o f learning disability. Journal of Learning Disabilities, 31, 67-82. Blagg, N . (1991). Can we teach intelligence? A comprehensive evaluation of Feuerstein's Instrumental Program. Hillsdale, N J : Erlbaum. Blatz, C . V . (1992). Contextual limits on reasoning and testing for critical thinking. In S.P. Norris (Ed.), The generalizability of critical thinking: Multiple perspectives on an educational ideal, (pp. 198-205). N e w York: Teachers College Press. Bradley, T . B . (1983). Remediation o f cognitive deficits: A critical appraisal o f the Feuerstein model. Journal of Mental Deficiency Research, 27, 79-92. Brandt, R. (1999). Educators need to know about the human brain. Phi Delta Kappan, 81, 235-238. Burden, R . (1987). Feuerstein's Instrumental Enrichment Programme: Important issues i n research and evaluation. European Journal of Psychology of Education, 11(1), 3-16. Campbell, D.T., & Stanley, J.C. (1963). Experimental and quasi-experimental designs for research. Chicago: Rand M c N a l l y . Cohen, J. (1988). Statistical power analysis for the behavioral sciences ( 2 ed.). Hillsdale, N J : Erlbaum n d  86  Cooper, H . (1998). Synthesizing Research: A guide for literature reviews. (3 ed.). Thousand Oaks, C A : Sage. r  Cronbach, L . J . , & Meehl, P.E. (1955). Construct validity i n psychological tests. Psychological Bulletin, 52, 281-302 Das, J.P., & Conway, R. (1992). Reflections on remediation and transfer: A Vygotskian perspective. In H . C . Haywood & D . Tzuriel (Eds.), Interactive assessment (pp. 94115). N e w Y o r k : Springer-Verlag. Dewey, J. (1933). How we think. Boston: D . C . Heath. Egan, G . (1998). The skilled helper: A problem-management approach to helping. Pacific Grove, C A : Brooks/Cole. Egozi, M . (1991). Instrumental Enrichment and mediation. In R. Feuerstein, P. K l e i n , & A.Tannenbaum (Eds.), Mediated learning experience (MLE): Theoretical, psychosocial and learning implications (pp. 347-364). London: Freund Publishing House. Emerson. L . W . (1991). M L E and American Indian education. In R. Feuerstein, P. K l e i n , & A.Tannenbaum (Eds.), Mediated learning experience (MLE): Theoretical, psychosocial and learning implications (pp. 133-156). London: Freund Publishing House. Ennis, R . H . (1985a). A logical basis for measuring critical thinking skills. Educational Leadership, 43, 44-48. Ennis, R . H . (1985b). Critical thinking and the curriculum. National Forum, 65, 28-31. Ennis, R . H . (1992). A concept o f critical thinking. Harvard Educational Review, 32, 81-111. Eysenck, J . H . (1978). A n exercise i n mega-silliness [Letter]. Psychological Bulletin, 85, 517. Eysenck, H . (1988). The concept o f "intelligence": Useful or useless? Intelligence, 12, 1-16. Feuerstein, R., & Feuerstein, S. (1991). Mediated learning experience: A theoretical review. In R. Feuerstein, P. K l e i n , & A.Tannenbaum (Eds.), Mediated learning experience (MLE): Psychosocial and learning implications (pp. 3 - 5 1 ) . London: Freund Publishing House. Feuerstein, R., Feuerstein, R., & Gross, S. (1997). The learning potential assessment device. In D . P . Flanagan, J.L. Genshaft & P . L . Harrison (Eds.), Contemporary intellectual assessment: Theories, tests, and issues (pp. 297-311). N e w Y o r k : Guilford Press. Feuerstein, R., & Hoffman, M . B . (1982). Intergenerational conflict o f rights: Cultural imposition and self-realization. Viewpoints in Teaching and Learning, 58, 44-63.  87  Feuerstein, R., Hoffman, M . B . , Rand, Y . , Jensen, M . R . , Tzuriel, D . , & Hoffman, D . B . (1986). Learning to learn: Mediated learning experiences and instrumental enrichment. Special Services in the Schools, 3, 49-82. Feuerstein, R., K l e i n , P. & Tannenbaum, A . (Eds.). (1991). Mediated learning experience (MLE): Psychosocial and learning implications. London: Freund Publishing House. Feuerstein, R., M i l l e r , R., Hoffman, M . B . , Rand, Y , Mintzker, Y . , & Jensen, M . R . (1981). Cognitive modifiability i n adolescence: Cognitive structure and the effects o f intervention. Journal of Special Education, 15, 269-287. Feuerstein, R., Rand, Y . , Haywood, H . C . , Hoffman, M . B . , & Jensen, M . R . (n.d.). L.P.A.D. Learning potential assessment device, experimental version. Jerusalem: HadassahWizo-Canada-Research Institute. Feuerstein, R., Rand, Y . , & Hoffman, M . B . (1979). The dynamic assessment of retarded performers: The Learning Potential Assessment Device: Theory, instruments and techniques. Baltimore: University Park Press. Feuerstein, R., Rand, Y . , Hoffman, M a l k a , Hoffman, Mendel, & M i l l e r , R . (1979). Cognitive modifiability i n retarded adolescents: Effects o f instrumental enrichment. American Journal of Mental Deficiency, 6, 519-550. Feuerstein, R., Rand, Y . , Hoffman, M . B . , & M i l l e r , R . (1980). Instrumental Enrichment: An intervention program for cognitive modifiability. Baltimore: University Park Press. Feuerstein, R., Rand, Y . , Jensen, M . R . , Kaniel, S., & Tzuriel, D . (1987). Prerequisites for assessment o f learning potential: The L P A D model. In. C S . L i d z (Ed.), Dynamic assessment (pp. 35-51). N e w York: Guilford. Fisher, R . A . (1932). Statistical methods for research workers (4 ed.). London: Oliver and Boyd. th  Frisby, C . L . , & Braden, J.P. (1992). Feuerstein's dynamic assessment approach: A semantic, logical, and empirical critique. Journal of Special Education, 3, 281-301. Gallo, P.S. (1978). Meta-analysis - a mixed meta-phor. American Psychologist, 33, 515-517. Glass, G . V . (1976). Primary, secondary, and meta-analysis o f research. Educational Researcher, 5, 3-8. Glass, G . V . (1977). Integrating findings: The meta-analysis o f research. In L . S . Shulman (Ed.), Review of research in education 5 (pp. 351-379). Itasca, I L : Peacock Publishers. Glass, G . V . , M c G a w , B . , & Smith, M . L . (1981). Meta-analysis in social research. Beverly Hills, C A : Sage.  88  Greenberg, K . H . , & Woodside, M . R . (1994). Differences in the degree o f mediated learning and classroom interaction structure for trained and untrained teachers. Journal of Classroom Interaction, 2(2), 1-9). Greenhouse, J.B., & Iyengar, S. (1994). Sensitivity analysis and diagnostics. In H . Cooper & L . V . Hedges (Eds.), The handbook of research synthesis (pp. 393-396). N e w Y o r k , N Y : Russell Sage. H a l l , J.A., Rosenthal, R., Tickle-Degnen, L . , & Mosteller, F. (1994). Hypotheses and problems in research synthesis. In H . Cooper & L . V . Hedges (Eds.), The handbook of research synthesis (pp. 17-28). N e w York, N Y : Russell Sage. H a m m i l l , D . D . (1993). A brief look at the learning disabilities movement i n the United States. Journal of Learning Disabilities, 26, 295-310. Haywood, H . C . (1977). A cognitive approach to the education o f retarded children. Peabody Journal of Education, 54, 110-116. Haywood, H . C . (1987). A mediational teaching style. The Thinking Teacher, 4(1), 1-6. Haywood, H . C , Arbitman-Smith, R., Bransford, J.D., Towery, J.R., Hannel, I.L., & Hannel, M . V . (1982). Cognitive education with adolescents: evaluation ofInstrumental Enrichment. Paper presented at Symposium o f 6th I . A . S . S . M . D . , Toronto. Haywood, H . C , Brown, A . L . , & Wingenfeld, S. (1990). Dynamic approaches to psychoeducational assessment. School Psychology Review, 19, AM-All. Haywood, H . C , Towery-Woolsey, J., Arbitman-Smith, R., & Aldridge, A . H . (1988). Cognitive education with deaf adolescents: Effects o f instrumental enrichment. Topics in Language Disorders, 8, 23-40. Haywood, H . C , Tzuriel, D . , & Vaught, S.R. (1992). Psychoeducational assessment from a transactional perspective. In H . C . Haywood & D . Tzuriel (Eds.), Interactive assessment (pp. 38-63). N e w York: Springer-Verlag. Hedges, L . V . (1981). Distribution theory for Glass's estimator o f effect size and related estimators. Journal of Educational Statistics, 6, 107-128. Hedges, L . V . (1982). Estimation o f effect sizes from a series o f independent experiments. Psychological Bulletin, 92, 490-499. Hedges, L . V . , (1994). Statistical considerations. In H . Cooper & L . V . Hedges (Eds.), The handbook of research synthesis (pp. 29-38). N e w York, N Y : Russell Sage. Hedges, L . V . (1986). Issues in meta-analysis. In E . Z . Rothkopf (Ed.), Review of research in education, N o . 13 (pp. 353-398). Washington, D C : American Educational Research Association.  89  Hedges, L . V . , & O l k i n , I. (1985). Statistical methods for meta-analysis. Toronto: Academic Press. Hunter, J.E., & Schmidt, F . L . (1990). Methods of meta-analysis: Correcting error and bias in research findings. Newbury Park, C A : Sage. Jensen, M . R . (n.d.). Cognitive modifiability and Instrumental Enrichment: A controlled evaluation of a classroom-based intervention model. Unpublished manuscript. Delphi Health & Science, and The National Center for Mediated Learning, Atlanta, G A . Jensen, M . , Feuerstein, R., Rand, Y . , Kaniel, S., & Tzuriel, D . (1988). Cultural difference and cultural deprivation: A theoretical framework for differential intervention. In R . M . Gupta & P. Coxhead (Eds.), Cultural diversity and learning efficiency (pp. 64-88). M a c M i l l a n Press. Johnson, R . H (1992). The problem o f defining critical thinking. In S.P. Norris (Ed.), The generalizability of critical thinking: Multiple perspectives on an educational ideal. (pp. 198-205). N e w Y o r k : Teachers College Press. Kaniel, S., & Feuerstein, R . (1989). Special needs o f children with learning difficulties. Oxford Review of Education, 15, 165-179. Kennedy, M . , Fisher, M . B . , & Ennis, R . H . (1991). Critical thinking: Literature review and needed research. In L . Idol & B . F . Jones (Eds.), Educational values and cognitive instruction: Implications for reform (pp. 11-40). Hillsdale, N J : Lawrence Erlbaum Associates. Kopp-Greenberg, K . H . (1991). A model for providing intensive mediated learning experience in preschool settings. In R. Feuerstein, P. K l e i n , & A.Tannenbaum (Eds.), Mediated learning experience (MLE): Psychosocial and learning implications (pp. 241-258).. London: Freund Publishing House. K u l i k , J . A . , & K u l i k , C - L . C . (1989). Meta-analysis i n education. International Journal of Educational Research, 13, 221-340. L ' A b b e , K . A . , Detsky, A . S , and O'Rourke, K . (1987). Meta-analysis i n clinical research. Annals of Internal Medicine, 107, 224-233. L i d z , C S . (1991). M L E components and their roots i n theory and research. In R . Feuerstein, P. K l e i n , & A.Tannenbaum (Eds.), Mediated learning experience (MLE): Psychosocial and learning implications (pp. 271-291). London: Freund Publishing House. Light, R.J., & Pillemer, D . B . (1984). Summing up: The science of reviewing research. Cambridge, M A : Harvard University Press. Link, F . R . (1991). Instrumental Enrichment. In A . L . Costa (Ed.), Developing minds: programs for teaching thinking: (Vol. 2 Rev. E d . , pp. 9-11). Alexandria, V A : Association for Supervision and Curriculum Development. 90  McPeck, J. (1981). Critical thinking and education. N e w York: St. Martin's Press. McPeck, J. (1990). Critical thinking and subject specificity: A reply to Ennis. Educational Researcher, 19(5), 10-12. McPeck, J . (1992). Thoughts on subject specificity. In S.P. Norris (Ed.), The generalizability of critical thinking: Multiple perspectives on an educational ideal, (pp. 198-205). N e w Y o r k : Teachers College Press. Marcia, J.E. (1966). Development and validation o f ego-identity status. Journal of Personality and Social Psychology, 3, 551-558. Marcia, J.E. (1980). Identity i n adolescence. In J. Adelson (Ed.), Handbook of adolescent psychology. N e w York: Wiley. Marcia, J.E., Waterman, A . S . , Matteson, D.R., Archer, S.L., & Orlofsky, J.L. (1993). Ego identiry: A handbook for psychosocial research. N e w York: Springer-Verlag. Missiuma, C , & Samuels, M . (1988). Dynamic assessment: Review and critique. Special Services in the Schools, 5(1-2), 1-22. Neisser, U . , Boodoo, G . , Bouchard, T., Boykin, A . W . , Brody, N . , Ceci, S., Halpern, D . , Loehlin, J., Perloff, R., Sternberg, R., & Urbina, S. (1996). Intelligence: Knowns and unknowns. American Psychologist, 51, 77-101. Norris, S.P. (1992). Introduction: The generalizability question. In S.P. Norris (Ed.), The generalizability of critical thinking: Multiple perspectives on an educational ideal. (pp. 1-15). N e w York: Teachers College Press. Orwin, R . G . (1983). A fail-safe ./V for effect size i n meta-analysis. Journal of Educational Statistics, 8, 157-159. Orwin, R . G . (1994). Evaluating coding decisions. In H . Cooper & L . V . Hedges (Eds.), The handbook of research synthesis (pp. 139-162). N e w York, N Y : Russell Sage. Perkins, D . N . , & Salomon, G . (1987). Transfer and teaching thinking. In D . N . Perkins, J. Lochhead & J . Bishop (Eds.), Thinking: The second international conference. Hillsdale, N J : Erlbaum Perkins, D . N . , & Salomon, G . (1994). In T. Husen, & T . N . Postlethwaite, (Eds.), The international encyclopedia of education (2nd ed., V o l s . 1-11, pp. 2916-2922). N e w Y o r k : Pergamon Perry, J.L. (1986). Field initiated research: A field based study of the effects of an educational program on school adjustment of mildly handicapped secondary school students. Final report. Raleigh, N C : Wake County Public School System. ( E R I C Document Reproduction Service N o . E D 312 848) 91  Petitti, D . B . (1994). Meta-analysis, decision analysis, and cost-effectiveness analysis. N e w Y o r k : Oxford University Press. Piaget, J. (1952). The origin of intelligence in children. N e w York: International Universities Press. Piaget, J. (1971). The theory o f stages i n cognitive development. In D . R . Green, M . P . Ford, & G . B . Flamer (Eds.), Measurement of Piaget. M c G r a w - H i l l . Piaget, J. (1985). The equilibrium of cognitive structures: The central problem of intellectual development (T. B r o w n & K . Julian Thampy, Trans.). Chicago: University o f Chicago Press. Presby, S. (1978). Overly broad categories observe important differences between therapies. American Psychologist, 33, 514-515. The Psychological Corporation. (1997). Wechsler Adult Intelligence Scale - Third Edition, Wechsler Memory Scale - Third Edition: Technical manual. Toronto, O N : Harcourt Brace. Rand, Y . , Tannenbaum, A . , & Feuerstein, R. (1979). Effects o f instrumental enrichment on the psychoeducational development o f low-functioning adolescents. Journal of Educational Psychology, 77,751-763. Rey, A . (1934). D'un procede pour evaluer l'educabilite (quelques applications en psychopathologie) [A procedure for assessing educability (some applications to psychopathology)]. Archives de Psychologie, 24, 297-337. Rey, A . (1962). Etude des insuffisances psychollgiques. II: La sysematisation des observations [The study o f psychological deficits. II: The systemization o f observations]. Neuchatell: Delachaux et Niestle. Rosenthal, R . (1979). The "file drawer" problem and tolerance for null results. Psychological Bulletin, 86, 638-641. Rosenthal, R . (1984). Meta-analytic procedures for social research. Newbury Park: Sage. Rosenthal, R . (1991). Meta-analytic procedures for social research: Applied Social Research Methods Series. ( V o l . 6). Newbury Park: Sage. Salomon, G . , & Perkins, D . N . (1989). Rocky roads to transfer: Rethinking mechanisms o f a neglected phenomenon. Educational Psychology, 81, 113-142. Samuels M . T., & Conte, R. (1987). Instrumental Enrichment with learning disabled adolescents: Is it effective? Journal of Practical Approaches to Developmental Handicap, 11, 4-6.  92  Samuels, M . , K l e i n , P., & Haywood, H . C . (1994) Cognitive modifiability. In T. Husen, & T . N . Postlethwaite, (Eds.), The international encyclopedia of education (2nd ed., V o l s . 1-11, pp. 857-861). N e w York: Pergamon Samuels, M . , K l e i n , P., & Haywood, H . C . (1995). Cognitive modifiability. Cognitive Education, 6(\), 3-7. Savell, J . M . , Twohig, P.T., & Rachford, D . L . (1986). Empirical status o f Feuerstein's "Instrumental Enrichment" (FIE) technique as a method o f teaching thinking skills. Review of Educational Research, 56, 381-409. Schmidt, F . L . (1992). What do data really mean? Research findings, meta-analysis, and cumulative knowledge i n psychology. American Psychologist, 10, 1173-1181. Secada, W . G . (1991) Student diversity and mathematics education reform. In L . Idol & B . F . Jones (Eds.), Educational values and cognitive instruction: Implications for reform (pp. 297-332). Hillsdale, N J : Lawrence Erlbaum Associates. Sewell, T., & Price, V . (1991). Mediated learning experience: Implications for achievement motivation and cognitive performance i n low socio-economic and minority children. In R. Feuerstein, P. K l e i n , & A.Tannenbaum (Eds.), Mediated learning experience (MLE): Psychosocial and learning implications (pp. 295-314). London: Freund Publishing House. Shayer, M . , & Beasley, F . (1987). Does instrumental enrichment work? British Educational Research Journal, 13, 101-119. Siegel, H . (1992). The generalizability o f critical thinking skills: Dispositions, and epistemology. In S.P. Norris (Ed.), The generalizability of critical thinking: Multiple perspectives on an educational ideal, (pp.97-108). N e w Y o r k : Teachers College Press. Slavin, R . E . (1986). Best-evidence synthesis: A n alternative to meta-analysis and traditional reviews. Educational Researcher, 15, 5-11. Stein, S.J., & Book, H . E . (2000). The EQ edge: Emotional intelligence and your success. Toronto: Stoddard Publishing. Sternberg, R . J . (1984). H o w can we teach intelligence? Educational Leadership, (September) 38-48. Sternberg, R . J . (1987). Questions and answers about the nature and teaching o f thinking skills. In J.B. Baron & R.J. Sternberg (Eds.), Teaching thinking skills: Theory and practice (pp. 251-259). N e w York: Freeman. Sternberg, R . J . (1994). Commentary: Reforming school reform: Comments on Multiple Intelligences: The theory i n practice. Teachers College Record, 95, 561-569.  93  Sternberg, R . J . (1996). Myths, countermyths, and truths about intelligence. Educational Researcher, 25, 11-16. Sternberg, R.J., & Bhana, K . (1986). Synthesis o f research on the effectiveness o f intellectual skills programs: Snake-oil remedies or miracle cures? Educational Leadership, 44, 6067. Stock, W . A . (1994). Systematic coding for research synthesis. In H . Cooper & L . V . Hedges (Eds.), The handbook of research synthesis (pp. 125-138). N e w York, N Y : Russell Sage. Sutton, A . J . , Duval, S.J., Tweedie, K . R . , Abrams, K . R . , & Jones, D . R . (2000). Empirical assessment o f effect o f publication bias on meta-analyses. British Medical Journal, 320, 1574-1577. Swartz, R.J. (1987). Teaching and thinking: A developmental model for the infusion o f thinking skills into mainstream instruction. In J.B. Baron & R.J. Sternberg (Eds.), Teaching thinking skills: Theory and practice (pp. 106-126). N e w Y o r k : W . H . Freeman. Tzuriel, D . (1991). Cognitive modifieability, mediated learning experience and affectivemotivational processes: A transactional Approach. In R. Feuerstein, P. K l e i n , & A.Tannenbaum (Eds.), Mediated learning experience (MLE): Psychosocial and learning implications (pp. 95-120). London: Freund Publishing House. Tzuriel, D . , & Alfassi, M . (1994). Cognitive and motivational modifiability as a function o f the instrumental enrichment (IE) program. Special Services in the Schools, 8, 91-128. Tzuriel, D . , & Haywood, H . C . (1992). The development o f interactive-dynamic approaches to assessment o f learning potential. In H . C . Haywood & D . Tzuriel (Eds.), Interactive assessment (pp. 3-37). N e w York: Springer-Verlag. . Vancouver School District (n.d.). Feuerstein's Instrumental Enrichment Curriculum Training Handbook. Vancouver, B C : Author. Vygotsky, L . S . (1962). Thought and language. Cambridge, M A : M I T Press. Vygotsky, L . S . (1978). Mind in society: The development of higher psychological processes. ( M . Cole, V . John-Steiner, S. Scribner, & E . Souberman Eds., and trans.). Cambridge, M A : Harvard University Press. Walter, G . A . , & Marks, S.E. (1981). Experiental learning and change: Theory design and practice. Toronto: Wiley. Wechsler, D . (1944). The measurement of adult intelligence ( 3 ed.). Baltimore: Williams & Wilkins. rd  94  Wechsler, D . (1975). Intelligence defined and undefined: A relativistic appraisal. American Psychologist, 30, 135-139. Wilgosh, L . , & Mulcahy, R . (1993). Cognitive educational models o f assessment, programming and instruction for native learners. Journal of Native Education, 20, 129-135. Wilkinson, L . , & Task Force on Statistical Inference. (1999). Statistical methods i n psychology journals: Guidelines and Explanations. American Psychologist, 54, 594604. Wolf, F . M . (1986). Meta-analysis: Quantitative methods for research synthesis. Beverly H i l l s : Sage. Wolfe, P., & Brandt, R . (1998). What do we know from brain research? Educational Leadership, 56, 8-13. Yates, S. (1987). Instrumental enrichment: A strategy for modifying cognitive ability. Australian Journal of Special Education, 11, 15-27.  95  Appendix A Summary of Studies Used, Measures and Effect Sizes Arbitman-Smith, R., & Haywood, H . C . (1980). Cognitive education for learningdisabled adolescents. Journal of Abnormal Child Psychology, 8, 51-64. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Mixed 10-11, approximately 5,6 Mixed Learning disabled Mixed 1 hour a day for 60 hours. Less than 1 year. Four instruments. Unknown which were used Pre-Post, Treatment Control Not classroom teacher Regular curriculum  A l l fifth- and sixth-grade students were rank-ordered on two IQ tests and were alternately assigned to either treatment or control groups. Data were reported only for those measures that showed a significant, positive, effect for the treatment group (IE). Information regarding teacher training was not reported. Dependent Measures: > Hiskey-Nebraska Test o f Learning Aptitude (Hiskey, 1966). Spatial Reasoning subtest. o Picture Identification subtest. Not used because F-test results were nonsignificant and data were not reported, o Ten items from the Verbal Analogies Test (Schlesinger and Shalom, Israel). N o t used because it was an informal measure, no differences were found between IE and control groups, and data were not reported, o Raven's Standard Progressive Matrices. Not used because analysis o f the pre- and post intervention scores failed to demonstrate differential gains between the two groups. Data were not reported, o Test o f Social Inference (Edmonson, de Jung, Leland, & Leach, 1974). Ten items were selected. Not used because it was an informal measure and total scores did not reveal a significant difference between the two groups. N o data were reported for the total scores. Effect Sizes Test  E.S.  8  Hiskey-Nebraska, Spatial Reasoning  Effect Size (Unbiased d)  9 Confidence intervals are set at 0.95  Confidence Confidence Interval - Minus Interval - Plus  Unable to calculate as only F-statistics were  0.7362 8  2-tailed p  96  reported.  9  Blagg, N . (1991). Can we teach intelligence? A comprehensive evaluation of Feuerstein's Instrumental Enrichment program. Hillsdale, N J : Erlbaum.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  British, mixed Not reported 14-16 Not reported, likely secondary level L o w average ability, low achievement Lower Attaining Pupil Program ( L A P P ) Mixed 2 years. 30-192 hours. Average: 112 hours 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization, 6. Instructions, 7. Temporal Relations, 8. Numerical Progressions, 9. Family Relations, 10. Illustrations, 16. Cartoons Pre-Post, Treatment Control Not classroom teacher Regular curriculum  The study took place i n four schools as a result o f a government-based initiative. Groups were chosen by teachers according to ability. There were no significant differences between control and experimental groups on pretest measures. Results that were analyzed by analyses o f covariance where sex and school factors were the covariates. Only significant results were reported for the teachers' assessment on the 16PF. The twelve teachers were trained, five o f them extensively.  Dependent Measures: > > > > >  o o  Cattell's 16 Personality Factor Questionnaire (16PF) - This measure was administered to the teachers. British A b i l i t y Scales ( B A S ) - Intelligence Quotient C S M S Science Reasoning Tasks Edinburgh Reading Test Richmond Attainment Test Battery > W o r k Study skills (RW-2) - Reading graphs and tables. > Work Study skills (RW-3) - Knowledge and use o f reference materials o W o r k Study skills (RW-1) - M a p reading. Not used as data were not suitable for calculation > Mathematics skills (RM-2) - Problem Solving o Mathematics skills (RM-1) - Concepts. Not used as data were not suitable for calculation. Teachers' observation o f pupils' behaviour schedule - Not used because it was an informal, locally developed measure Teachers' observation o f pupils' attitudes - Not used because it was an informal, locally developed measure 97  o o o  Coopersmith Self-Esteem Scale (Anglicized) - Not used because W i l c o x o n Sign Rank scores were graphed. It was not possible to analyze the results, Pupils' attitudes toward school work - Not used because informal, locally developed rating scales were used, Three Semantic-Differential Scales administered to teachers: M y s e l f as a Teacher, The Characteristics o f Less Academic Pupils, Attitudes Towards Instrumental Enrichment N o t used because it was an informal, locally developed measure.  Effect Sizes Test  E.S.  B A S IQ Score 0.0438 Edinburgh Reading Test 0.0435 Richmond R M 2 - Math Problem Solving 0.3520 CSMS Science Reasoning 0.1602 Richmond RW2 - Reference Materials 0.2403 Richmond RW3 - Reference Materials 0.0394 16PF - Teachers (Low = Humble, High = Assertive) 16PF - Teachers (Low = Tough-Minded, High = Tender-Minded) 16PF - Teachers (Low = Trusting, High = Suspicious)  1.2270 -1.0661 0.9876  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only t-statistics were reported. Unable to calculate as only t-statistics were reported. Unable to calculate as only t-statistics were reported.  Brainin, S.S. (1982). The effects of Instrumental Enrichment on the reasoning abilities,  reading achievement and task orientation of sixth grade underachievers. Unpublished doctoral dissertation, Columbia University, N e w York.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Low 11 - 12 Grade 6 Low Underachievers Mixed 1 year, 59 hours 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, and one-third o f 4. Analytic Perception Pre-Post, Treatment Control Not classroom teacher Regular curriculum 98  Students were randomly assigned to "small core" classes. Treatment and control teachers were selected without regard to their class composition. Treatment and Control groups were independent. Tests analyzed were good ones. The two experimental teachers received approximately 30 hours o f training during the school year. In addition, the investigator, an educational psychologist trained and experienced i n IE, paid ten visits to the experimental classrooms during the course o f the study observing and assisting. Dependent Measures: > Thorndike-Hagen Cognitive Abilities Test, Level C , Non-Verbal Battery. > Comprehensive Test o f Basic Skills ( C T B S ) - Total reading score. o Criterion-Referenced Test ( C R T ) - Not used because it was locally developed and criterion-referenced. o Devereux Elementary School Behavior Rating Scales ( D E S B R S ) - Not used because only gain scores were reported. Effect Sizes Test Thorndike-Hagen, Nonverbal Comprehensive TBS - Reading  E.S. 0.3221 -0.3877  2-tailed p 0.2670 0.2380  Confidence Confidence Interval - Minus Interval - Plus -0.2444 0.8886 -1.0330 0.2578  Bruce, J., & Martin, D . (1984). Cognitive improvement of hearing-impaired high school students through instruction in Instrumental Enrichment. Working paper presented at the International Symposium on Cognition, Education, and Deafness, Gallaudet College, Washington, D C . ( E R I C Document Reproduction Service N o . E D 247 725) Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  American, mixed Caucasian and Black Not stated Average for IE: 16.4, average for control: 16.7 Secondary Mixed Hearing Impaired M a i n l y female 1 year Four instruments: parts-wholes, comparison (3. Comparisons), projection o f virtual relationships, spatial relationships (2. Orientation in Space I) Pre-Post, Treatment Control Not classroom teacher Regular curriculum  Experimental and control groups were matched on the bases o f sex, age and level o f class placement (remedial, regular or advanced placement). Experimental teachers completed an intensive six-day training session.  99  Dependent Measures: > Raven's Progressive Matrices - abstract reasoning, non-verbal > K i t o f Factor Referenced Cognitive Test ( K F R C T ) - Diagramming Relationships and Letter Sets - abstract reasoning, "verbal" o Written problem solutions - Not used because it was locally developed o Teacher observation checklist - Not used because it was locally developed o Stanford Achievement Test, Hearing Impaired Version ( S A T - H I ) - Reading Comprehension, Math Concepts and Math Computation - Not used because data were not reported Effect Sizes Test KFRCT - Diagramming Relationships  E.S. 0.1981  KFRCT - Letter Sets 0.2526 Raven's Progressive Matrices 0.5128  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported  Church, S. (1994). A longitudinal study of the effects of Feuerstein's Instrumental Enrichment on intersensory integration, academic achievement and referral for special education programs. Unpublished doctoral dissertation, University o f Oklahoma, Norman. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Mixed Approximately 9 at the beginning o f the study 4 L o w reading and gifted Special Education students Mixed 1 year, 64 hours plus 2, 3 and 4 year follow-up testing 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception Pre-Post, Treatment Control, Longitudinal testing at the end o f grades 4, 5, 6, and 7 Not classroom teacher Regular curriculum  A pre-post design with longitudinal measurement. The treatment group had mainly special education students. Control group was compared with N C E scores from the previous year and no significant difference was found between the two groups. Tests were good ones. Instruction was conducted by a trained IE instructor on an itinerant basis. Dependent Measures: > Bender Gestalt Visual Motor Integration Test. Elizabeth Koppitz (1958) standardized developmental age scoring system was used. 100  o  > >  Scores were not analyzed at the end o f the first year because they were for only the experimental group with no control comparison. Visual Aural Digit Span ( V A D S ) Test (Koppitz, 1977). Iowa Test o f Basic Skills (ITBS). Effect Sizes  End of Treatment Year Test Bender Gestalt Visual Motor Integration  E.S. 1.0581  Visual Aural Digit Span (VADS) Iowa Test of Basic Skills  0.5440 -0.0645  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. -0.5624 0.4335 0.8026  Follow-up 1 Year After Treatment Bender Gestalt Visual Aural Digit Span Iowa Test of Basic Skills  1.6775 1.2051 0.3700  0.00001** 0.0002** 0.2340  0.9743 0.5477 -0.2400  2.3808 1.8626 0.9800  Follow-up 2 Years After Treatment Iowa Test of Basic Skills  0.3584  0.2502  -0.2513  0.9681  Follow-up 3 Years After Treatment Iowa Test of Basic Skills  0.8894  0.0062**  0.2553  1.5235  Eichler, P.J.S. (1994). A study of a cognitive developmental intervention for troubled or troubling adolescents. Unpublished doctoral dissertation, Peabody College o f Vanderbilt University, Nashville, T N . Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed L o w to Middle 14-18 Institute o f Learning Research (ILR) L o w to Average Seriously Emotionally Disturbed (SED), Behaviour Disordered (BD), Learning Disabled ( L D ) Mixed 4 - 6 weeks 4. Analytic Perception Pre-Post, Treatment Control Not classroom teacher - Experimenter Regular curriculum  Students were paired by learning disabilities, sex, age, race and IQ, and then randomly divided into two groups. One was designated as control and the other as experimental. Statistical analysis was analyzed using the Wilcoxon Matched-Pairs Signed-Ranks test. However, raw 101  scores were reported for each group and it was possible to calculate means and standard deviations. The experimenter was well trained i n the use o f the L P A D and IE teaching methods. The researcher taught IE and administered the tests. Dependent Measures: > Watson-Glaser Critical Thinking Appraisal, Form A ( C T A - A ) (1966, 1977, 1980) Assesses cognitive functions. > Learning Potential Assessment Device ( L P A D ) (Feuerstein, 1979) - Set Variations I and Representational Stencil Design tests ( R S T D ) > Picture Motivation Scale ( P M S ) (Kunca & Haywood, 1969) - Intrinsic motivation > The M a z e Task (Delclos & Haywood. Reported i n Haywood & Switzky, 1986) Intrinsic Motivation Effect Sizes Test LP AD - Set Variations I LPAD - RSTD Watson-Glasser Critical Thinking Picture Motivation Scale Haywood's Maze Test  E.S.  0.5441 2.4590 0.9668 0.4201 2.2468  2-tailed p 0.2340 0.00001** 0.0404* 0.3524 0.0000**  Confidence Confidence Interval - Minus Interval - Plus -0.3485 1.4367 3.6204 1.2975 0.0405 1.8931 1.3063 -0.4661 1.1274 3.3663s  Feuerstein, R., Rand, Y . , Hoffman, M . , & Miller, R. (1980). Instrumental Enrichment: An intervention program for cognitive modifiability. Glenview, I L : Scott, Foresman. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  Israeli and African and Asian immigrants Low 12 - 15 at the beginning o f the study Residential and Day Centres - Secondary Levels Low Mixed Predominantly male 2 years, 5 hours per week 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization, 6. Instructions, 7. Temporal Relations, 8. Numerical Progressions, 9. Family Relations, 10. Illustrations, 11. Syllogisms (to some classes as the end), 12. Representational Stencil Design, 13. Orientation i n Space II, Pre-Post, Treatment Control Not classroom teacher General Enrichment  The control group received "General Enrichment," that is, enriched and supplementary input o f content and curriculum-oriented experiences. Treatment and control students had low school performance and behavioural problems. Treatment and control groups were matched 102  pairs and were independent. The quality o f tests is unknown. Over the two years teachers received approximately 27 days o f training.. In addition, training was continued in-service by the supervisor at each research site. The supervisor visited each IE class once weekly, intervening upon request, and offering guidance based on the post-lesson critique. Dependent Measures: (Data used were combined residential centre and day centre results) > Primary Mental Abilities ( P M A ) test, version 4 to 6 (Thurstone, 1965) - Total score > Project Achievement Battery. General knowledge, nature, geography, reading comprehension. o Antonyms was not used because reading comprehension was considered a better example o f reading ability. Geometry, addition, subtraction, multiplication and division were not used because they were considered to be too specific examples o f arithmetic to fit into the analysis. Part-Whole and Bible were also not used because they did not fit into this analysis. > Classroom Participation Scale (CPS I & II) (Tannenbaum & Levine, 1968). Six factors from factor analyses, the first four required an effect size with a minus sign to indicate positive growth. The other two were self-sufficiency and adaptiveness to work demands. > Embedded Figures Test (EFT) (Witkin, 1950). Total Correct. The subject must identify a given part within the whole. o Average Time was not used. > Human Figure Drawing ( H F D ) (Witkin et al., 1962). The subject draws figures o f both sexes and the drawings are scored for sophistication and detail. > Kuhlmann-Finch Postures Test (Finch, 1953), modified. A test o f spatial orientation. A human figure is presented next to similar but rotated figures. The subject must choose the figure identical to the model. > Lahy Test (Zazzo, 1964). Proportion Correct. Figures are created by the rotation o f a protruding bar from the angles and midpoints o f a small square. o Total number attended and Proportion wrong were not used. > Terman Nonverbal Intelligence Test (Terman, 1942). Concept formation tasks. The subject must select a geometric design that does not conform to the grouping principles in a group o f geometric designs. > D-48 Test (Gough & Domino, 1963). Involves analogies and progressions i n relationships among dominoes. > Porteus M a z e Test (Porteus, 1965). o Levidal Self-Concept Scale (Levine & Katz, 1971). Data were not reported. See Rand, Tannenbaum & Feuerstein (1979) for that data o Dapar Intelligence Test - Not used because it was not possible to combine residential centre and day centre results, o Hebrew Language Development - Not used because it was not possible to combine residential centre and day centre results, o L P A D Set Variations - Not used because results were reported i n terms o f observed deficient cognitive functions, and because reported results were the difference between pre- and post-test changes.  103  Effect Sizes Embedded Figures Human Figure Drawing Lahy Test Terman Concept Formation-Nonverbal D-48 Porteus Maze Kuhlman-Finch Postures, modified CPS I, Adaptiveness to Work Demands CPS II, Self Sufficiency PMA - Total Score PAB - General Knowledge PAB - Reading PAB-Nature PAB - Geography  E.S. 0.7974 -0.4417 0.4744 0.4764 0.9011 0.1400 0.7014 0.8480 0.9388 0.5307 0.3195 0.0177 -0.0374 0.1712  2-tailed p 0.0000** 0.0198* 0.0124** 0.0120** 0.0000** 0.4532 0.0004** 0.0000** 0.0000** 0.0054** 0.0892 0.9282 0.8414 0.3628  Lower d scores are best below CPS I - Acting Out CPS I - Unsocialized Behaviour CPS I - Immaturity CPS -1 Interpersonal Conduct  -0.1796 -0.0710 -0.0886 -0.3827  0.3370 0.7040 0.6384 0.0434*  Test  Confidence Confidence Interval - Minus Interval - Plus 0.4159 1.1788 -0.8133 -0.8133 0.1022 0.8467 0.8487 0.1041 1.2864 0.5158 -0.2276 0.5076 1.0796 0.3231 1.2312 0.4646 0.5520 1.3256 0.1572 0.9043 0.6890 -0.0499 -0.3494 0.3849 0.3298 -0.4046 -0.1966 0.5391  -0.5475 -0.4382 -0.4560 -0.7532  0.1883 0.2963 0.2787 -0.0122  Funk, P . G . (1987). Modifying the cognitive functioning of the learner: The effects of two years' instruction in Instrumental Enrichment (Feuerstein). Unpublished doctoral dissertation, Temple University, Philadelphia, P A .  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Mixed 12-14 6-8 Average - L o w reading and math Learning Disability resource room Mixed 2 years, 30 minutes every other day. Unknown Pre-Post, Treatment Control Unknown Students i n the same remedial program, but i n a different school (Regular curriculum).  Control and experimental groups had similar pre-test Reading Comprehension, Total Mathematics and IQ score scores. Tests were good ones. Information on training o f IE teachers was unavailable.  104  Dependent Measures: > >  Stanford Achievement Tests, Advanced Level, Form E - Total Mathematics and Reading Comprehension Otis-Lennon School Ability Test, Intermediate Level, Form R - IQ scores  Effect Sizes Test  E.S.  Otis-Lennon SAT - Total Math SAT - Reading Comprehension  0.9657 0.9367 0.5043  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.0004** 0.0010** 0.0588  0.4218 0.3945 -0.0186  1.5100 1.4789 1.0271  Genasci, K . (1983). The effects of Instrumental Enrichment on aptitudes and affective measures of adolescent students in selected classroom settings. Unpublished doctoral dissertation, University o f Oregon, Eugene, O R .  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  American, mixed Mixed 1 1 - 1 5 and 1 5 - 1 8 7 - 8 and students i n alternative high school classrooms H i g h and M i x e d Mixed Not reported - M i x e d 1 year 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception Pre-Post, Treatment Control Two were not the classroom teacher and one was. Regular curriculum  Design: Instructor: Control Group:  Eight school districts were asked to participate in the study. Three were able to do so, and they provided the experimental and control group classrooms. Groups were both selected by the school districts and randomly drawn from classrooms. Groups that were chosen were done so on the basis o f their similarity to the randomly selected groups. Teachers had received one academic year o f IE training. Results that were used i n the meta-analysis were from A N C O V A F-tests with the 7 and 8 grade students. th  th  Dependent Measures: >  o  Academic Self-Concept Measure (Payne, 1962). Part one: The students' perception o f him/herself as a student. Part two: Students rate how they think teachers have perceived them as students. Data used in the meta-analysis were from the high achieving grade 7 - 8 , age 1 1 - 1 5 groups only. Primary Mental Abilities test ( P M A ) (Thurstone, 1965). N o t used because the results were calculated with the Mann-Whitney U non-parametric statistical procedure and data were not reported in a fashion that could be used to calculate effect sizes.  105  Effect Sizes Test Academic Self-Concept - Student's Perceptions of Themselves Academic Self-Concept - Student's Perceptions of How Teachers Rate Them  E.S. 0.0253 0.0270  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported.  Graham, E . E . (1981). Feuerstein's Instrumental Enrichment used to change cognitive and verbal behaviour in a city-core, multi-ethnic Toronto secondary school. Unpublished doctoral dissertation, University o f Toronto, Toronto, O N . Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  Canadian, multi-ethnic Average-low Approximately 14 Nine L o w English (Six classes divided into treatment and control groups) Common and remedial English classes Mixed 1 year. (7 A months); three 45-minute periods weekly 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 6. Instructions, 10. Illustrations Pre-Post, Treatment Control. Incomplete Nested design Not classroom teacher Regular English curriculum l  The study used an incomplete nested design. The composition o f all grade nine classes was randomly determined. Good tests. Descriptive statistics were not reported so effect sizes were calculated from A N C O V A F tests. Experimental group teachers had 3 days o f training and the researcher wrote manuals for them for each o f the instruments used. In addition, peer coaches with expertise i n IE and M L E were used to help other students. Dependent Measures > Lorge-Thorndike Intelligence Test, Non-Verbal Batteries III and I V , Total score > Standard Diagnostic Reading Test, Brown Level, Reading Comprehension, Test 2, Total score. > Piers-Harris Self-Concept Scale, Total score. o Primary Mental Abilities Letter Series Test Levels 1 and II. N o t used because total scores were not reported, o A writing test devised by the researcher. Not used because it was an informal measure.  106  Effect Sizes Test  E.S.  Lorge-Thorndike III - Total 0.2779 Lorge-Thorndike IV - Total 0.2689 Stanford Diagnostic Reading-Total Score 0.0217 Piers-Harris - Total Score 0.0807  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate reported. Unable to calculate reported. Unable to calculate reported. Unable to calculate reported.  as only F-statistics were as only F-statistics were as only F-statistics were as only F-statistics were  H a l l , J . N . (1981). Evaluation and comparison: Social Learning Curriculum and Instrumental Enrichment. Final Report. Nashville, T N : George Peabody College for Teachers. ( E R I C Document Reproduction Service N o . E D 244 484.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design:  Instructor: Control Group:  American, mixed Mixed 11-17 5 - 9 , predominantly elementary M i x e d L o w and high ability Varying Exceptionalities: Learning Disabled ( L D ) , Educable Mentally Handicapped ( E M H ) , Behaviour Disordered (BD), Learning Problems ( L P ) M i x e d - predominantly male 1 year - 4 hours per week 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception Non-Equivalent Control Group, Pre-Post - Treatment Control There were three groups, Social Learning Curriculum ( S L C ) , IE, and control. S L C data were not used i n this analysis. Classroom teacher Traditional special education programs (Regular curriculum)  This was a Non-Equivalent Control Group design with two treatment groups and a control group. Students in both treatment and control groups had learning problems. IE teachers attended training workshops on four occasions during the year-long treatment; however, information on how long the workshops lasted was not reported. Good tests.  Dependent Measures: > >  Standard Progressive Matrices ( S P M ) (Raven, 1960) - intellectual functioning, nonverbal Test o f Social Inference (TSI) (Edmonson, Leland, deJung, & Leach, 1974) inferential thinking 107  >  > > o  Matching Familiar Figures Test ( M F F T ) (Kagan, Rosman, Day, Albert, & Phillips, 1964), number correct - nonverbal problem solving - reflectivity-impulsivity, and visual-perceptual acuity Peabody Individual Achievement Test (PIAT) (Dunn & Markwardt, 1970) - General Information subtest. Piers-Harris Children's Self-Concept Scale (PH) (Piers, 1969) Test o f the Hierarchy o f Social Knowledge ( T H I N K ) (Smith & Greenberg, n.d.) - Not used because it was used in a pilot fashion with only some o f the subjects.  Effect Sizes Test Raven's Standard Matrices Matching Familiar Figures, # correct PIAT - General Information Piers Harris Children's Self-Concept Test of Social Inference  E.S. 0.3542 0.6477 0.4759 -0.0323 0.1627  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.1118 0.0042** 0.0332* 0.8808 0.4592  -0.0805 0.2057 0.0390 -0.4639 -0.2700  0.7890 1.0900 0.9131 0.3993 0.5950  Haywood, H . C . , Towery-Woolsey, J., Arbitman-Smith, R., & Aldridge, A . H . (1988). Cognitive education with deaf adolescents. Effects of Instrumental Enrichment. Topics in Language Disorders, 92, 23-40.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  American, mixed Mixed 12-16 7-9 Mixed Deaf Mixed 1 Vi academic years/85 - 170 hours A l l students received: 1. Organization o f Dots, 2. Orientation in Space I, 4. Analytic Perception. Students with the most hours also received: 3. Comparisons, 7. Temporal Relations, 8. Numerical Progressions, 9. Family Relations, 13. Classifications, 15. Orientation i n Space II Pre-Post, Treatment Control Classroom teacher Regular curriculum  Effect sizes were calculated from F-tests as standard deviations were not reported. Tests were good ones. Teachers were trained (the length of training was not reported), and an experienced teacher-trainer made regular consulting visits to the school to monitor progress and provide help.  Dependent Measures: >  Raven's Standard Progressive Matrices ( R S P M ) 108  > > > > >  o  Primary Mental Abilities test ( P M A ) - Reasoning subtest, Spatial Relations subtest K e y M a t h Diagnostic Test (Connolly, Nachtman, & Pritchett, 1976). The M a t h Reasoning subtest was chosen for this analysis. Peabody Individual Achievement Test (PIAT) (Dunn & Markwardt, 1970). Math, Reading and Information subtests. Metropolitan Achievement Test (Durost, Bixler, Hildreth, Lund, & Wrightstone, 1959). Reading, Math, Language, Science, Social Studies subtests. Stanford Achievement Tests (Gardner, Rudman, Karlsen, & M e r w i n , 1984) ( S A T ) . Reading Comprehension, Language, Science, and the Math Applications subtest were chosen for this analysis. Mastery tests for the IE program - Not used because it was not a standardized measure, and the purpose o f this study is to evaluate the effectiveness o f F I E i n other areas.  Effect Sizes Test  E.S.  P M A - Reasoning 1.2477 Raven's Standard Progressive Matrices 0.6991 PIAT - Reading 0.1412 Metropolitan Achievement - Reading 0.1799 SAT - Reading Comprehension 0.0360 PIAT - Information 0.5150 Metropolitan - Language 0.24508 SAT - Language 0.5518 KeyMath - Math Reasoning 0.5376 PIAT - Math 0.1832 Metropolitan - Math 0.0538 SAT - Math Applications 0.0478 Metropolitan - Science 0.2557  SAT - Science 0.1465  Metropolitan - Social Studies 0.0087  109  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported.  Hirsch, J. (1987). A study of a program based on Feuerstein's theories intended to  teach high-level cognitive skills to African-American and Mexican-American Junior-HighSchool students identified as learning disabled. Unpublished doctoral dissertation, University of San Francisco, C A .  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  American. African and Mexican-American Working class 13 - 1 6 + , approximately 7-9 L o w achievement Learning Disabled. Students in resource specialist classes. Mainly male 1 year, 1 period per day 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization, 6. Instructions, 7. Temporal Relations, 8. Numerical Progressions Pre-Post, Treatment Control Not classroom teacher - The researcher? Regular remedial instruction  The study used a non-equivalent control group design. T w o Resource Specialist classes at the same junior-high school were compared. There was no reason to suspect differential recruitment for treatment. There were no significant differences between groups on pre-test measures o f English scholarship, Reading, Math, citizenship and absences. The researcher was the IE teacher. She had spent over ten years working with Feuerstein in Israel and as a resource teacher and trainer in the United States. She used IE and the L P A D teaching IE to students and adults.  Dependent Measures: > > > o  Wide Range Achievement Test - Math, and Reading Recognition (decoding) subtests. Brigance Diagnostic Inventory o f Basic Skills - Reading Comprehension subtest Comprehensive Test o f Basic Skills ( C T B S ) - Reading and Math subtests. English and Math scholarship, citizenship and absences. Not used because these were informal measures from report cards.  Effect Sizes Test W R A T - Math W R A T - Reading Recognition Comprehensive TBS - Math Brigance - Reading Comprehension Comprehensive TBS - Reading  E.S. 0.1115 0.0000 0.7426 0.1296 0.0460  110  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.7872 1.0000 0.0784 0.7490 0.9124  -0.6893 -0.8002 -0.0847 -0.6714 -0.7543  0.9123 0.8002 1.5699 0.9306 0.8463  Jensen, M . R . (n.d.). Cognitive modifiability and Instrumental Enrichment: A controlled evaluation of a classroom-based intervention model. Unpublished manuscript. Delphi Health & Science, and The National Center for Mediated Learning, Atlanta, G A . Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  American, mixed Low 13-14 (Average: 13.10) at the beginning (16-17) 7, 8 at the beginning Low Special education, inner-city M i x e d - mainly male 18 months (midpoint) and 36 months (endpoint) 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization, 6. Instructions, 7. Temporal Relations, 8. Numerical Progressions, 9. Family Relations, 10. Illustrations, 11. Transitive Relations and Syllogisms, 12. Representational Stencil Design, 13. Orientation i n Space II, 14. Classifications, 15. Organization o f Dots II Pre-Post, Treatment Control Not classroom teacher Regular curriculum  Experimental and control schools did not differ significantly i n terms o f gender or IQ. Testing personnel were blind to the specific hypotheses o f the study. Effect sizes were calculated from t- and F-tests as well as from means, standard deviations and number o f subjects. Teachers were trained (the length o f time was not reported) and received intermittent coaching assistance from experienced IE teachers. "The variable H i g h IE - L o w IE was created by dividing the sample at the median based on teacher estimates o f the amount o f IE students received" (p. 19). Dependent Measures: Data were reported according to an acquisition-retention, near transfer (resistanceflexibility) and far transfer (transformability-generalizability) model. Midpoint > Near transfer and far transfer tasks were selected from the French K i t (Educational Testing Service) or adapted from "up-coming and more advanced parts o f the IE curriculum." > Piers-Harris Children's Self-Concept Scale (Piers, 1984) > The Haywood Mazes. (Haywood has argued that this test may constitute a fairly good measure o f intrinsic motivation.) Endpoint. Near Transfer: > Primary Mental Abilities Test (Thurstone, 1965) - Spatial Relations, Figure Grouping, W o r d Grouping and Perceptual Speed subtests. > Raven's Standard Progressive Matrices (1977) - Series 3 - 5 .  111  Far Transfer: > > >  > o  o  Cognitive Abilities Test (Thorndike & Hagen, 1971) - Number Series (48 items) Familiar Words Questionnaire (Haywood, 1981) A mathematics test designed especially for the project. Although this was an informal measure and should have been deleted from this analysis, it was not possible to isolate it and do so. Piers-Harris Children's Self-Concept Scale (Piers, 1984) Acquisition. N o t used because measures were represented by IE materials covered by the intervention and adapted to a test format. This analysis is calculating the effects IE has on other domains. Teacher Likert-type ratings o f their own students. N o t used because it was an informal measure.  Effect Sizes Midpoint Test Far Transfer, Sample A Far Transfer, Sample B Far Transfer, Sample C Near Transfer, Sample A Near Transfer, Sample B Near Transfer, Sample C Haywood Maze Test, Samples A & B  E.S. 0.1621 0.8171 0.1581 0.2665 0.9154 0.4866 0.4437  Haywood Maze Test, Sample C 0.5521 Piers-Harris, Sample A 0.6718 Piers-Harris, Sample B 1.3263 Piers-Harris, Sample C 0.1004  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.5486 -0.3645 0.6890 0.0054** 1.3939 0.2403 0.6342 0.5156 -0.3180 -0.2791 0.8120 0.3370 1.5132 0.3176 0.0026** 0.0043 0.9690 0.0478* Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported.  Endpoint Test Far Transfer, High IE  Unbiased d 0.1484  Far Transfer, Low IE -0.2781 Near Transfer, High IE 0.3508 Near Transfer, Low IE -0.4120  112  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only t-statistics were reported. Unable to calculate as only t-statistics were reported. Unable to calculate as only t-statistics were reported. Unable to calculate as only t-statistics were reported.  Kenney, M . V . (1984). Effects of Feuerstein's Instrumental Enrichment on the reasoning, non-verbal, intelligence, and locus of control of 12 to 15 year old educable mentally handicapped and learning disabled students. Unpublished doctoral dissertation, University o f Missouri-Columbia, Columbia, M O .  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Mixed 12-15 Six - nine Low E M H and L D Mixed 1 year, 125 hours 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization, 8. Numerical Progressions Pre-Post, Treatment Control, Not classroom teacher Regular curriculum  Most students were randomly selected from their classrooms. Treatment and Control groups were independent. Tests were good. Twenty-five volunteer teachers received five, 9-hour days o f training from an experienced IE teacher and trainer. Twelve were randomly selected from this group to teach the IE group.  Dependent Measures: >  > >  Woodcock-Johnson Psycho-Educational Battery: Reasoning Cluster (Woodcock & Johnson, 1977). Four subtests: Analysis-Synthesis, Concept Formation, Analogies and Antonyms-Synonyms. Standard Progressive Matrices (Raven, 1977) Intellectual Achievement Responsibility Questionnaire (Crandall, Katkovsky & Crandall, 1965)  Effect Sizes Test Raven's Standard Matrices (EMH) Raven's Standard Matrices (LD) Woodcock-Johnson Reasoning (EMH) Woodcock Johnson Reasoning (LD) Intellectual Achievement Responsibility Questionnaire, Total I scores (LD) Intellectual Achievement Responsibility Questionnaire, Total I scores (EMH)  E.S. 0.0901 0.3150 0.3800 0.7937  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.6514 0.7566 -0.4713 0.7951. 0.1970 -0.1652 0.9461 0.1868 -0.1861 0.2982 1.2893 0.0010**  0.7130  0.0046**  0.2210  1.2051  0.1088  0.7040  -0.4527  0.6702  113  McRainey, L . (1982). Social outcomes of a cognitive education program. Unpublished doctoral dissertation, Vanderbilt University, Nashville, T N . Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: F I E Instruments Used: Design: Instructor: Control Group:  American, mixed Mixed 12-13 7,8 Mixed Mixed Mixed 1 semester, 20 - 23 hours Not reported Pre-Post, Treatment Control Not classroom teacher Creative Dramatics class  The treatment group consisted o f students who chose IE as an enrichment course that was offered, among other courses, as part o f the regular curriculum. The control group consisted o f students who had selected Creative Dramatics as their enrichment course. The group was chosen on the basis o f class size, makeup, format and willingness o f the teachers to participate. Pre-test results from the Raven's did not reveal any significant differences between the two groups. The F-test was used to calculate the effect size because standard deviations were not reported. The IE teacher received her training at a workshop held during the year before the study. Dependent Measures: > Classroom Environment Scale (Trickett & Moos, 1974) - assesses the social climate i n the classroom. Students respond to forced-choice items. M a i n effects o f groups across all subscales and trials were examined i n this meta-analysis, o Raven's Standard Progressive Matrices (Raven, 1958) - Not used because it was administered as a pre-test measure o f the comparability o f the two groups, o Teacher's Rating Form - Not used as it was locally developed, o Reciprocal Category System (Ober, 1970) - A measure o f verbal interactions. Not used because data were examined from the framework o f single subject research, and the data were reported i n anecdotes and graphs. Effect Sizes  Test  E.S.  Classroom Environment Scale 0.6716  114  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported.  Martin, D . S . (1984). Cognitive modification for the hearing impaired adolescent: The promise. Exceptional Children, 51, 235-242.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Mixed Average: 16.5 at the beginning. Secondary L o w reading Hearing impaired Mixed 2 years 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 6. Instructions, 14. Classifications Pre-Post, Treatment Control. Pilot project Unknown Regular curriculum  This is a pilot study with hearing impaired students. Treatment and Control groups were matched on the basis o f age, sex, degree o f hearing loss and reading ability. Raven's test is a good one. Information on teacher training was not reported.  Dependent Measures: > o  o o  Raven's Standard Progressive Matrices. Stanford Achievement Test-Hearing Impaired Version ( S A T - H I ) . Reading subtest: Not used because data were not given, only a report o f the results within groups, not between. The Math subtest results were not used because the number o f experimental subjects was nine and that number was compared to the total student body o f 253. Problem-Solving Interviews. Not used because they were informal measures. Teacher Observation o f Behaviour. Not used because the reported results were prepost within the experimental group only.  Effect Sizes Test  E.S.  Raven's Standard Progressive Matrices 0.8963  115  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported.  M a x c y , P. (1991) The effects of Feuerstein's Instrumental Enrichment (FIE) on cognitive performance and transfer measures of at-risk adolescents when adequate mediated learning experiences are present. Unpublished master's thesis, Simon Fraser University, Burnaby, British Columbia, Canada. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  Canadian, mixed M i x e d (low to middle income) 12-14 Eight (Bridge) M i x e d - L o w achievement At-risk adolescents with emotional and social difficulties Mixed 8 months 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization (parts), 6. Instructions (parts) Pre-Post, Treatment Control Not classroom teacher Regular curriculum  Experimental and control subjects were all students in an at-risk "bridging" program. IE was an elective course for experimental students. The IE teacher (the researcher) was very knowledgeable about F I E theory, F I E curricula and the L P A D . Other experienced IE teachers made periodic, unannounced visits and filled out a rating form. The tests were good ones. Dependent Measures: > Raven's Standard Progressive Matrices (Raven, Court & Raven, 1983) > Test o f Cognitive Skills ( C T B / M c G r a w - H i l l , 1 9 8 1 ) - Total score > Intellectual Achievement Responsibility Scale (Crandall, 1983) > Coopersmith Self-Esteem Questionnaire (Coopersmith, 1987) o Canadian Achievement Test ( C T B / M c G r a w - H i l l , 1982) - not used because it was administered to only the treatment group o Learning Potential Assessment Device, Group Organization o f Dots Test, Group Numerical Progressions, Group Representational Stencil Design Test - not used because it was administered to only the treatment group Effect Sizes Test Raven's Standard Matrices Test of Cognitive Skills Intellectual Achievement Responsibility Coopersmith  E.S. 0.3039 0.4536 0.9215 0.2344  116  2-tailed p 0.5418 0.3734 0.0750 0.6892  Confidence Confidence Interval - Minus Interval - Plus 1.2751 -0.6675 1.4538 -0.5465 1.9358 -0.0929 -0.9284 1.3972  Mulcahy, R. (1993). Cognitive education project (Report N o . ISBN-0-7732-1147-0). Alberta Deptartment o f Education, Edmonton, A B . ( E R I C Document Reproduction Service N o . E D 367 679)  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  Canadian, mixed Mixed Approximately 9 and 12 at the beginning Grades 4 and 7 at the beginning Mixed Gifted, Average, and Learning Disabled Mixed 1 year, 2 years, Maintenance year Year 1 : 1. Organization o f Dots, 2. Orientation in Space I, 3. Comparisons, 4. Analytic Perception Year 2: 6. Instructions, 7. Temporal Relations, 8. Numerical Progressions, 9. Family Relations, 10. Illustrations, , 1 3 . Classifications. Pre-Post, Treatment Control Classroom teacher The control group received regular curriculum instruction. There was another comparison group, S.P.E.L.T., that was not analyzed i n this meta-analysis.  A Pre-post factorial design was used. A t pre-test approximately 4000 students were screened for possible inclusion i n three diagnostic categories (gifted, average and learning disabled). Eighty percent o f the teachers who volunteered were randomly assigned to the three conditions (I.E., S.P.E.L.T., Control). The remaining teachers were assigned to the control group. Students were selected according to results on the C A T , the C C A T and the S R B G C S S . Results were analyzed and at grade 4 no significant differences were observed for the conditions. Grade 7 IE students were rated as being somewhat higher i n motivation than control students. In-service training for IE was conducted for 5 full days. Standardized tests were good ones. This study was conducted over 3 years with 3 time periods for measurement o f affect and achievement: pre-test (end o f year 1), post-test (end o f year 2) and another post-test at the end of 1 year o f maintenance (p.96). The authors reported all means and standard deviations, and the number o f subjects for only the end o f year 1 (pp. 87 and 91). The analysis for the other two measurement points consisted o f calculating F-tests for three, not two, groups. Therefore, this meta-analysis calculated effect sizes only for the measures used at the end o f the first year. The data were taken from pages 87, 91 and 247-268.  Dependent Measures: >  >  Canadian Achievement Test ( C A T ) ( M c G r a w - H i l l Ryerson, 1983). Used for selection as well as evaluation. Reading Vocabulary, Reading Comprehension, Math Calculation and Math Concept Application subtests. Harter's Perceived Competence Scale (Harter, 1982) - Children's views o f their academic abilities (cognitive), social skills and popularity, physical abilities and general self-esteem. 117  > > o  o o o o  o  o  Coopersmith Self-Esteem Inventory (Coopersmith, 1 9 8 1 ) - attitudes toward self i n social, academic, family and personal areas. Intellectual Achievement Responsibility Questionnaire (Crandall, Katkovsky, & Crandall, 1965) - locus o f control for outcomes in academic achievement situations. The Scales for the Rating o f the Behavioral Characteristics o f Superior Students ( S R B G C S S ) (Renzulli & Hartman, 1976). - Not used, as it was a pre-test measure used for selection. Canadian Cognitive Abilities Test ( C C A T ) (Thorndike & Hagen, 1982) - N o t used, as it was a pre-test measure used for selection, Reading Awareness Questionnaire (Paris & Oka, 1986) - Not used as a measure o f reading ability. Cloze Task. Passages selected from the Spache Diagnostic Reading Test (1972). - N o t used, as it was locally developed, Error Detection Task (Comprehension Monitoring). Stories were selected from textbooks and from the Spache Diagnostic Reading Test (1972). - Not used, as it was locally developed. Perceived Problem Solving Inventory (PPSI) (Heppner & Petersen, 1982) - examines the underlying dimensions o f people's perceptions o f their real-life, personal problemsolving process. Used at the final test point. Not used as only data at the end o f the first year were analyzed. M a t h Problem Solving Strategy Assessment. To assess the types o f strategies students used. - Not used, as it was locally developed.  Effect Sizes Age 9-10 Test CAT - Math Computation (LD) CAT - Math Concept Application (LD) CAT - Math Computation (Average) CAT - Math Concept Application (Average) CAT - Math Computation (Gifted) CAT - Math Concept Application (Gifted) CAT - Reading Vocabulary (LD) CAT - Reading Comprehension (LD) CAT - Reading Vocabulary (Average) CAT - Reading Comprehension (Average) CAT - Reading Vocabulary (Gifted) CAT - Reading Comprehension (Gifted) Harter's Perceived Competence Academic (LD) Harter's Perceived Competence Academic (Average) Harter's Perceived Competence Academic (Gifted) Coopersmith - School (LD) Coopersmith - School (Average)  E.S. 0.1052 0.4800 0.0582  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.6818 -0.3967 0.6071 0.0644 0.9886 -0.0285 0.8026 -0.3949 0.5113  0.2179 -0.0398  0.3472 0.8650  -0.2364 -0.5092  0.6723 0.4295  -0.0251 0.2100 -0.1061 -0.1238  0.9204 0.4122 0.6818 0.5892  -0.4944 -0.2933 -0.6080 -0.5772  0.4442 0.7125 0.3958 0.3297  0.0939 -0.2692 -0.1552  0.6818 0.2628 0.5156  -0.3594 -0.7406 -0.6252  0.5471 0.2022 0.3148  -0.5714  0.0286*  -1.0828  -0.0600  0.6725  0.0046**  0.2068  1.1381  0.3766 -0.8305 0.1969  0.1188 0.0032** 0.3954  -0.0968 -1.3526 -0.2572  0.8500 -0.3084 0.6510  118  Coopersmith - School (Gifted) Intellectual Achievement Responsibility Questionnaire (LD) Intellectual Achievement Responsibility Questionnaire (Average) Intellectual Achievement Responsibility Questionnaire (Gifted) Harter - Physical Abilities (LD) Harter - Physical Abilities (Average) Harter - Physical Abilities (Gifted) Harter - Social Skills & Popularity (LD) Harter - Social Skills & Popularity (Average) Harter - Social Skills & Popularity (Gifted) Harter - General Self-Esteem (LD) Harter - General Self-Esteem (Average) Harter - General Self-Esteem (Gifted) Coopersmith - General (LD) Coopersmith - General (Average) Coopersmith - General (Gifted)  0.0550  0.8180  -0.4144  0.5244  -0.3141  0.2224  -0.8187  0.1905  -0.1371  0.5552  -0.5906  0.3170  -0.2186 -0.2504 0.4263 0.1796 -0.6852  0.3628 0.3320 0.0688 0.4532 0.0094**  -0.6893 -0.7538 -0.0318 -0.2906 -1.2009  0.2521 0.2531 0.8844 0.6404 -0.1696  0.4086  0.8020  -0.0491  0.8663  0.3330 -0.5475 0.1430 0.3733 -0.9372 0.1243 0.2391  0.1676 0.0358* 0.5352 0.1212 0.0004** 0.5892 0.3222  -0.1396 -1.0581 -0.3106 -0.1000 -1.4648 -0.3291 -0.2319  0.8055 -0.0368 0.5966 0.8467 -0.4096 0.5778 0.7109  E.S. 0.0758 0.0717 0.5443  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.7642 -0.4234 0.5750 0.9124 -0.4275 0.5708 0.0444* 1.0757 0.0128  Age 12-13 Test CAT - Math Computation (LD) CAT - Math Concept Application (LD) CAT - Math Computation (Average) CAT - Math Concept Application (Average) CAT - Math Computation (Gifted) CAT - Math Concept Application (Gifted) CAT -Reading Vocabulary (LD) CAT -Reading Comprehension (LD) CAT -Reading Vocabulary (Average) CAT -Reading Comprehension (Average) CAT -Reading Vocabulary (Gifted) CAT - Reading Comprehension (Gifted) Harter's Perceived Competence Academic (LD) Harter's Perceived Competence Academic (Average) Harter's Perceived Competence Academic (Gifted) Coopersmith - School (LD) Coopersmith - School (Average) Coopersmith - School (Gifted) Intellectual Achievement Responsibility Questionnaire (LD)  -0.4686 0.2547  0.0818 0.4066  -0.9978 -0.3450  0.0606 0.8543  -0.0848 -0.0938 -0.2033 -0.2988 0.1962 0.0051 -0.0521  0.7794 0.7114 0.4238 0.2670 0.4654 0.9840 0.8650  -0.6827 -0.5931 -0.7036 -0.8241 -0.3276 -0.5926 -0.6498  0.5131 0.4055 0.2970 0.2270 0.7200 0.6027 0.5467  0.6872  0.0088**  0.1740  1.2004  -0.1053  0.6966  -0.6282  0.4177  -0.0456 0.0391 0.3789 -0.0235  0.8808 0.8808 0.1586 0.9362  -0.6433 -0.4600 -0.1480 -0.6212  0.5521 0.5381 0.9058 0.5742  -0.1962  0.4412  -0.6964  0.3039  119  Intellectual Achievement Responsibility Questionnaire (Average) Intellectual Achievement Responsibility Questionnaire (Gifted) Harter - Physical Abilities (LD) Harter - Physical Abilities (Average) Harter - Physical Abilities (Gifted) Harter - Social Skills & Popularity (LD) Harter - Social Skills & Popularity (Average) Harter - Social Skills & Popularity (Gifted) Harter - General Self-Esteem (LD) Harter - General Self-Esteem (Average) Harter - General Self-Esteem (Gifted) Coopersmith - General (LD) Coopersmith - General (Average) Coopersmith - General (Gifted)  -0.2108  0.4296  -0.7347  0.3132  0.4297 0.1149 0.1151 -0.0961 0.3250  0.1616 0.6528 0.6672 0.7566 0.2040  -0.1736 -0.3845 -0.4085 -0.6940 -0.1772  1.0329 0.6143 0.6375 0.5019 0.8272  0.4742  0.7840  -0.0551  1.0036  -0.3269 -0.1806 0.6349 -0.6195 -0.3481 0.3534 0.0083  0.2262 0.4778 0.0198* 0.0466 0.1738 0.1868 0.7872  -0.8552 -0.6806 0.1003 -1.2287 -0.8508 -0.1730 -0.5894  0.2015 0.3194 1.1700 -0.0103 0.1546 0.8797 0.6060  Muttart, Kathy. (1984). Assessment o f effects o f Instrumental Enrichment cognitive  training. Special Education in Canada, 58, 106-108. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  Canadian, mixed Mixed 5-12 Seven, eight Low Students in remedial programs Mixed 9 months 1. Organization o f Dots, 2. Orientation i n Space I, 4. Analytic Perception Pre-Post, Treatment Control Not reported Regular curriculum  Information was not reported on pupil or IE teacher selection procedures, or on the teacher's level o f training. Tests were good ones.  Dependent Measures: > > >  Primary Mental Abilities Test ( P M A ) , Composite score (Thurstone, 1963) - Full-scale ability Canadian Test o f Basic Skills ( C T B S ) , Composite score ( K i n g & Hieronymus, 1975) General and social knowledge Brookover Self-Concept o f Ability Scale (Brookover, Patterson & Thomas, 1962) Self-concept 120  > > >  St. John Academic Self-Concept Scale (St. John, 1971) - Self-concept Achievement Self-Esteem Scale (Danley, 1980) - Self-esteem Self-Concept Scale (Lipsett, 1958)  Effect Sizes Test  P M A Composite Canadian TBS Composite St. John Academic Self-Concept Achievement Self-Esteem Brookover Self-Concept of Ability Lipsett Self-Concept Scale  E.S.  2-tailed p  0.8946 0.9738 -0.4934 0.4094 0.0160 1.1733  0.0784 0.0574 0.3174 0.4066 0.9760 0.0258*  Confidence Interval - Minus  -0.1042 -0.0333 -1.4601 -0.5529 -0.9364 0.1425  Confidence Interval - Plus  1.8933 1.9809 0.4733 1.3717 0.9684 2.2042  Pasquier, D . (1991). Evaluation des effets d'un montage pedagogique educabilite M P E - sur la modifiabilite structurelle cognitive - M S C - d'adultes en formation: une experience en formation de macons. / Effects o f an "educability" program on cognitive structural modifiability i n adults: A study o f masons i n training. European Review of Applied Psychology, 41, 125-136.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  French Unknown Adults Technical school students Unknown. Mixed? Subjects were being trained as masons Male Four 1-hour sessions per week. 1 year or 1 semester. 1. Organization o f Dots, 8. Numerical Progressions, 11. Transitive Relations and Syllogisms, 12. Representational Stencil Design Pre-Post, Treatment Control Not classroom teacher Traditional instruction/training with classical pedagogic methods  Selection procedures were not reported. Information on teachers' training was not reported. Tests are unknown. Treatment was for four 1 -hour sessions per week, but it was not reported i f it was for 1 year or 1 semester.  Dependent Measures: > o  S P E 129 - A test o f intellectual level; it measures concrete logic, visual logic, propositions and numerical sequences, and gives a score o f "general intelligence. A modification o f the Test du Raisonnement (Reasoning Test) (Schircks, A . , 1970). Not used because it was locally developed.  121  Effect Sizes Test SPE 129  E.S. 1.5121  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 2.5419 0.0040** 0.4823  Perry, J.L. (1986). Field initiated research: A field based study of the effects of an educational program on school adjustment of mildly handicapped secondary school students. Final report. Raleigh, N C : Wake County Public School System. ( E R I C Document Reproduction Service N o . E D 312 848)  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, M i x e d Mixed Approximately 1 2 - 1 5 7 and 8 - secondary Mixed Varying Exceptionalities: Learning Disabled ( L D ) , Educable Mentally Handicapped ( E M H ) , Emotionally Handicapped (EH) Predominantly male Grade seven - 2 years, grade eight - 1 year, and retest Not reported Pre-Post, Treatment Control, Random assignment to groups Not reported Regular curriculum  Students i n special education classes i n selected schools were randomly assigned to treatment or control conditions. This resulted in groups having balanced exceptionalities. Information on teacher training was not reported. Tests were good ones.  Dependent Measures: > > >  o o  Otis-Lennon School A b i l i t y Test - used as a measure o f verbal-educational intelligence. Peabody Picture Vocabulary - used as a measure o f language, i.e., receptive vocabulary Burks Behaviour Rating Scale. Subtests: Excessive Self-Blame, Excessive Anxiety, Excessive Withdrawal, Excessive Dependency, Poor Ego Strength, Poor Physical Strength, Poor Coordination, Poor Intellectuality, Poor Academics, Poor Attendance, Poor Impulse Control, Poor Reality Control, Poor Sense o f Identity, Excessive Suffering, Poor Anger Control, Excessive Sense o f Persecution, Excessive Aggression, Excessive Resistance, Poor Social Conformity. These subtests required an effect size with a minus sign to indicate positive growth. Wechsler Intelligence Scale for Children (WISC) - Not used because it was used as a pre-test only. Activity Participation Measure, grade nine - Not used as these were not standardized measures 122  o  Weighted-grade-point-average ( W G P A ) , school attendance ( A B S ) , special education students meeting the academic course requirement through taking non-special education academic courses ( A C A D ) and the number o f course hours taken for the total academic year (HRS). Grades seven, eight and nine. - N o t used as these were not standardized measures  Effect Sizes 1 Year Test Otis Lennon School Ability PPVT Lower d scores are best below Burks - Excessive Aggression Burks - Excessive Self Blame Burks - Excessive Anxiety Burks - Excessive Withdrawal Burks - Excessive Dependency Burks - Poor Ego Strength Burks - Poor Physical Strength Burks - Poor Coordination Burks - Poor Intellectuality Burks - Poor Academics Burks - Poor Attendance Burks - Poor Impulse Control Burks - Poor Reality Control Burks - Poor Sense of Identity Burks - Excessive Suffering Burks - Poor Anger Control Burks - Excessive Sense of Persecution Burks - Excessive Resistance Burks - Poor Social Conformity  E.S. 0.0830 0.2524  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.5666 0.7530 -0.4007 0.2758 -0.2003 0.7050 -0.4842 -0.2678 -0.2185 -0.2713 -0.2274 -0.1682 -0.4024 0.0117 -0.2621 -0.1142 -0.2662 -0.4559 -0.2181 -0.1896 -0.3171 -0.5220 -0.3726 -0.4264 -0.2506  0.4178 0.6360 0.6863 0.6324 0.6773 0.7380 0.4997 0.9251 0.6418 0.7938 0.6376 0.4461 0.6868 0.7160 0.5858 0.3802 0.5297 0.4757 0.6535  -0.0332 0.1841 0.2339 0.1805 0.2250 0.2849 0.0487 0.4684 0.1898 0.3400 0.1857 -0.0049 0.2344 0.2632 0.1343 -0.0709 0.0786 0.0246 0.2015  0.8886 0.4238 0.3124 0.4354 0.3320 0.2186 0.8336 0.0444* 0.4122 0.1416 0.4180 0.9840 0.3078 0.2542 0.5620 0.7566 0.7338 0.9124 0.3844  Otis-Lennon PPVT  E.S. -0.0197 0.0495  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.9204 -0.4211 0.3818 0.4451 0.8026 -0.3461  Lower d scores are best below Burks - Poor Social Conformity Burks - Excessive Aggression Burks - Excessive Self Blame Burks - Excessive Anxiety Burks - Excessive Withdrawal Burks - Excessive Dependency Burks - Poor Ego Strength Burks - Poor Physical Strength Burks - Poor Coordination  0.1259 0.1912 -0.0468 0.0086 0.1939 -0.2613 0.0544 -0.2393 -0.1008  2 Years Test  123  0.5286 0.3422 0.8180 0.9680 0.3370 0.1970 0.7872 0.2380 0.6170  -0.2699 -0.2052 -0.4424 -0.3869 -0.2024 -0.6583 -0.3413 -0.6361 -0.4965  0.5218 0.5875 0.3489 0.4042 0.5903 0.1358 0.4500 0.1576 0.2950  Burks Burks Burks Burks Burks Burks Burks Burks Burks Burks -  Poor Intellectuality Poor Academics Poor Attendance Poor Impulse Control Poor Reality Control Poor Sense of Identity Excessive Suffering Poor Anger Control Excessive Sense of Persecution Excessive Resistance  -0.2177 -0.3009 -0.0760 -0.1242 0.0895 0.2857 0.0888 -0.0061 -0.2345 0.0322  0.2802 0.1388 0.7040 0.5352 0.6600 0.1586 0.6600 0.9760 0.2460 0.8728  -0.6143 -0.6990 -0.4717 -0.5201 -0.3062 -0.1116 -0.3069 -0.4016 -0.6313 -0.3633  0.1789 0.0967 0.3197 0.2716 0.4852 0.6831 0.4845 0.3895 0.1622 0.4278  Rand, Y . , Mintzker, Y . , Miller, R., Hoffman, M . B : , & Friedlender, Y . (1981). The Instrumental Enrichment program: Immediate and long-term effects. In P. Mittler (Ed.), Frontiers of knowledge in mental retardation: Vol. 1. Social, educational, and behavioural aspects (pp. 141-152). Baltimore, M D : University Park Press.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  Israeli and African and Asian immigrants Low 16-19 Draftees into the Israeli Defence Forces Low Mixed Predominantly male Follow-up 2 years after the original Instrumental Enrichment study. See Rand et al. (1979) and Feuerstein etal. (1980) None Pre-Post, Treatment Control Not classroom teacher General Enrichment  This was a follow-up measurement 2 years after the original study by Feuerstein et al. (1980). Comparisons were between 57 matched pairs as well as on a larger population that was preand post-tested. The quality o f tests is unknown.  Dependent Measures: > > > o  Dapar Intelligence Test. A composite score derived from an Otis-type verbal test and a nonverbal test based on Raven's Progressive Matrices. Hebrew Language Development Test. A composite score from reading, writing, comprehension and vocabulary measures. Zadakh. A composite score is obtained by a structured interview. A general prediction rating is given regarding future army success. K a b a - Not used because, in addition to using the educational background o f the subject, it is a composite o f the three tests, above.  124  Effect Sizes Test Dapar Intelligence Test Hebrew Language Development Zadakh - Army Success Prediction  E.S. 0.4475 -0.0383 -0.2090  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.1571 0.0026** .0.7379 0.2371 0.7872 -0.3138 0.2460 -0.5637 0.1456  Rand, Y . , Tannenbaum, A . J., & Feuerstein, R. (1979). Effects o f Instrumental Enrichment on the psychoeducational development o f low-functioning adolescents. Journal of  Educational Psychology, 71, 751-763. This paper presents the results o f the study which Feuerstein, Rand, Hoffman & M i l l e r (1980), above, detail. It gives data on the Levidal Self-Concept Scale that was omitted from the 1980 publication.  Effect Sizes Test Levidal - Failure at School Levidal - Motivation for Learning Levidal - Confidence in Success  E.S. -0.0868 -0.0315 0.0673  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus -0.4541 0.2805 0.6456 0.3356 -0.3987 0.8650 -0.3000 0.4345 0.7188  Robinson, S.S. (1983). The effectiveness of Instrumental Enrichment in changing the learning and behavior problems of emotionally disturbed children and adolescents. Unpublished doctoral dissertation, Vanderbilt University, Nashville, T N .  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  American, mixed Mixed 10, 11 - ? Elementary Low Emotionally disturbed M i x e d (six boys, three girls) 12 weeks, 50 hours 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 10. Illustrations (selected pages to four children) Pre-Post, Treatment Control Not classroom teacher Regular curriculum  Control group students were matched to the distribution o f the IE students. Treatment and Control groups were independent. The teacher had taught IE to emotionally disturbed children 125  before, but was not a member o f the school staff. She was initially unfamiliar with the students. Dependent Measures: > Raven's Standard Progressive Matrices > Daily Behavior Rating Scale - developed by the U C L A / N P I Parent Training C l i n i c . There were results for ratings by parents and for ratings by teacher-counsellors. o Mastery Test 1 - N o t used because it was a measure o f the mastery o f IE material only, o Vocabulary Test - N o t used because it was a measure o f familiarity with vocabulary taught in the first two instruments o f the IE program, o M o o s Family Environment Scale - Not used because group data were not reported. Results were graphed individually, o Semi-Structured Interview - Not used because information was reported in narrative form for each individual child, o File Data Sheet - N o t used because information was presented i n narrative form for each child. Effect Sizes Test Raven's Progressive Matrices, Standard Daily Behaviour Rating Scale - Parent Ratings Daily Behaviour Rating Scale - Teacher Ratings  E.S. -0.2199 -0.6428 0.6142  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.6384 0.7068 -1.1467 Unable to calculate as only t-statistics were reported. Unable to calculate as only t-statistics were reported.  Rusling, E . G . (1984). An exploration of the initial effects of the first two instruments of Instrumental Enrichment. Unpublished doctoral dissertation, University o f Rochester, Rochester, N Y . Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Middle class 12-14 Junior H i g h school Class A : L o w average, Class B : Handicapped to Borderline L D and E M H Mixed 5 months 1. Organization o f Dots, 2. Orientation i n Space I Pre-Post, Treatment Control. Nonequivalent control group design. Not classroom teacher Regular curriculum  The study used a Nonequivalent control group design in which intact classrooms are used and were assumed to represent collectives. Treatment was nested under classrooms. The subjects 126  themselves did not select treatment or nontreatment. Treatment and Control groups were independent. Over the period o f the study teachers received a total o f six days o f training on the two IE instruments used. A support person was also trained and assigned to the classes. Tests were good ones.  Dependent Measures: > > > > > > >  Motor-Free Visual Perception Test ( M V P T ) (Colarusso & Hammil, 1972) Benton Visual Retention Test ( B V R T ) (Benton, 1974) (Visual-motor) Porteus Mazes A g e Score (Porteus, 1955) (Visual-motor) Trail-Making Test, Part A (Reitan, 1971) (Visual-motor, numbers) Trail-Making Test, Part B (Reitan, 1971) (Visual-motor, numbers and letters) Raven Coloured Progressive Matrices, Sets A , A B , B (Raven, 1947) Raven Standard Progressive Matrices, Sets C , D , E  Effect Sizes Test Raven's Coloured Matrices, A, AB, B Class A Raven's Coloured Matrices, A, AB, B Class B Ravens Standard Matrices, C, D, E, Class A Ravens Standard Matrices, C, D, E, Class B Motor-Free Visual Perception, Class A Motor-Free Visual Perception, Class B Porteus Maze, Class A Porteus Maze, Class B Lower d scores are best below Benton error score, Class A Benton error score, Class B Trail-Making, Part A, Time, Class A Trail-Making, Part A, Time, Class B Trail-Making, Part B, Time, Class A Trail-Making, Part B, Time, Class B  E.S.  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus  -0.3200  0.4654  -1.1818  0.5419  0.3934  0.3734  -0.4712  1.2580  0.1361  0.7566  -0.7213  0.9935  0.4552 -0.2095 0.8082 0.0000 1.3566  0.3030 0.6312 0.0750 1.0000 0.0052**  -0.4122 -1.0683 -0.0824 -0.8564 0.4070  1.3226 0.6492 1.6988 0.8564 2.3061  0.1250 -0.5811 0.8907 0.1581 0.3825 -0.1349  0.7718 0.0124* 0.0524 0.7188 0.3344 0.7566  -0.7322 -1.4553 -0.0071 -0.6996 -0.4817 -0.9922  0.9822 0.2932 1.7884 1.0158 1.2466 0.7225  127  Samuels, M . , Roadhouse, A . , Conte, R., & Zirk, H . (1984). Instrumental enrichment with low achieving adolescents. Alberta Department o f Education. ( E R I C Document Reproduction Service N o . E D 263 244)  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  Canadian, M i x e d Mixed 12-17 Secondary (Vocational School) Low Mixed Mixed 1 year, 2 years - approximately 270 hours, maintenance year(year 3) Year One: 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization, 8. Numerical Progressions, 10. Illustrations, Year 2: 1. Organization o f Dots (continued), 6. Instructions, 7. Temporal Relations, 8. Numerical Progressions (continued), 9. Family Relations, 10. Illustrations Pre-Post, Treatment Control with follow-up test i n year 3 Not classroom teacher Regular curriculum  Students were selected on the basis o f having a F u l l Scale IQ o f at least 90, below average achievement, no history o f serious attendance or behaviour problems and no history o f language difficulties. H a l f o f this group was randomly assigned to experimental and control groups. Treatment and Control groups were independent within each year, but not among the 3 years. Testing revealed no significant differences between the groups on any variable. The experimental group was taught by an itinerant teacher from the Calgary Learning Centre who had received three, 1 -week long workshops on the theory, goals, and intervention techniques of IE. The teacher was also observed and supervised by one o f the authors o f the study. Tests were good ones.  Dependent Measures: > > > >  >  Raven's Standard Progressive Matrices (Raven, 1956) Wechsler Intelligence Scale for Children-Revised ( W I S C - R ) - Used for year 2 Woodcock-Johnson Tests o f Cognitive Ability (Woodcock, R . W . , & Johnson, M . B . , 1977) - Reasoning Cluster - Not used for year 2 as the W I S C - R was used. Learning Potential Assessment Device ( L P A D ) (Feuerstein, Rand & Hoffman, 1979, 1980) > Numerical Progressions - Year 3 > Set Variations II - Years 2 and 3 > R e y ' s Complex Figure Drawing Test - Year 2 Canadian Tests o f Basic Skills 128  >  >  >  o o o  > Reading Vocabulary > Reading Comprehension > Total M a t h Woodcock-Johnson Tests o f Achievement (Woodcock, R . W . , & Johnson, M . B . ) > Reading Cluster > Mathematics Cluster Holtzman and Brown Survey o f Study Habits and Attitudes and Williams Perception o f Thinking Abilities Scale - a combined modified version o f these tests was used. Eight subtests: Attitude Towards Education, Attitude Towards Teacher, Study Habits & Test Taking, Motivation, Academic Interest, Need Achievement, Achievement Anxiety ( A n effect size with a minus sign indicates positive growth), Perception o f Thinking A b i l i t y Stony Brook Classroom Observation Code. Eight subtests, four o f which required an effect size with a minus sign to indicate positive growth: Interference, Off-Task Behaviour, M i n o r Motor Behaviour and Gross Motor Behaviour. The other four were Absence o f Behaviours, Responding to Questions, Solicitation, Request Clarification Stanford Diagnostic Mathematics Test - Not used in as it was administered only i n the initial test period before treatment, Teacher Behaviour Ratings - Not used as these were not standardized measures Attendance and Attrition Records - Not used as these were not standardized measures  Effect Sizes 1 Year Test LP AD Numerical Progressions H & B Attitude Towards Education H & B Attitude Toward Teacher H & B Study Habits & Test Taking H & B Motivation H & B Academic Interest H & B Need Achievement H & B Perception of Thinking Ability Stony Brook, Absence of Behaviour Stony Brook, Responds to Questions Stony Brook, Solicitation Stony Brook, Request for Clarification Woodcock-Johnson Reasoning Woodcock-Johnson Math Cluster Canadian TBS - Total Math Woodcock-Johnson Reading Cluster Canadian TBS - Reading Vocabulary Canadian TBS - Reading Comprehension  E.S. 0.7917 -1.1987 -0.4278 -1.0422 -0.8313 -1.1962 0.5015 -1.0441 0.3363 -0.2331 -0.2436 0.6158 -0.1444 -0.7431. -0.4568 0.0668 -0.5716 -0.2835  Lower d scores are best below H & B Achievement Anxiety Stony Brook - Interference Stony Brook-Off Task Stony Brook - Minor Motor Movement Stony Brook - Gross Motor Standing  0.0683 -0.2449 -0.3588 -0.5159 -0.2972 129  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 1.9238 0.1706 -0.3405 -0.0149 0.0478* -2.3825 0.6750 0.4472 -1.5306 0.0784 -2.2039 0.1195 0.3050 0.1528 -1.9675 -0.0128 0.0478* -2.3796 0.3734 1.6088 -0.6059 0.0784 -2.2061 0.1178 0.4066 -0.4589 1.1315 0.5592 0.5620 -1.0255 -1.0362 0.5490 0.5486 0.1362 -0.1922 1.4237 0.7948 -1.2363 0.9475 0.3841 0.1970 -1.8704 -1.5612 0.6477 0.4180 0.9044 -1.0240 1.1576 0.3124 0.5408 -1.6839 0.8124 0.6100 -1.3793  0.4532 0.5418 0.3788 0.2076 0.2327  -1.0224 -1.0375 -1.1547 -1.3184 -1.0912  1.1591 0.5478 0.4372 0.2867 0.4968  2 Years Test LP AD Set Variations II WISC-R Full Scale Score WISC-R Performance Scale WISC-R Verbal Scale H & B Attitude Towards Education H & B Attitude Toward Teacher H & B Study Habits & Test Taking H & B Motivation H & B Academic Interest H & B Need Achievement H & B Perception of Thinking Ability Stony Brook, Absence of Behaviour Stony Brook, Responds to Questions Stony Brook, Solicitation Stony Brook, Request for Clarification WISC-R Full Scale Woodcock-Johnson Reasoning Woodcock-Johnson Math Cluster Canadian TBS - Total Math Woodcock-Johnson Reading Cluster Canadian TBS - Reading Vocabulary Canadian TBS - Reading Comprehension  E.S. 0.5659 -0.2342 -0.2797 -0.1804 -1.0131 -0.0311 0.4148 -0.3729 -0.0117 -0.0128 -0.8148 -0.1733 -0.6621 -0.1813 0.2616 -0.2342 -0.2868 -0.5222 -1.6784 0.1528 -0.0332 -0.0820  Lower d scores are best below H & B Achievement Anxiety Stony Brook - Interference Stony Brook-Off Task Stony Brook - Minor Motor Movement Stony Brook - Gross Motor Standing  -0.0546 -0.0540 0.7864 -0.5705 0.8172  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.3174 -0.5460 1.6778 0.6694 0.6100 -1.1378 0.6252 0.5418 -1.1846 0.7220 0.6966 -1.0828 0.1448 0.0872 -2.1710 1.0594 0.9522 -1.1216 0.4592 -0.6872 1.5169 0.5092 -1.4727 0.7269 1.0788 0.9840 -1.1022 1.0777 0.9840 -1.1033 0.3197 0.1616 -1.9493 0.7093 0.7040 -1.0559 0.2425 0.1528 -1.5666 0.7015 0.6892 -1.0640 1.1462 0.5620 -0.6231 0.6694 -1.1378 0.6100 0.8092 0.6100 -1.3828 0.5865 0.3576 -1.6310 -0.4114 0.0094** -2.9455 1.2448 0.7872 -0.9392 0.9522 ' -1.1237 1.0573 -1.1722 1.0089 0.8808 0.9204 0.0904 0.0910 0.2150 0.0802  -1.1453 -0.9352 -0.1277 -1.4690 -0.0994  1.0360 0.8271 1.7004 0.3280 1.7338  3-Year Follow-up Test LPAD Numerical Progressions LPAD Set Variations II Raven's Standard Matrices H & B Attitude Towards Education H & B Attitude Toward Teacher H & B Study Habits & Test Taking H & B Motivation H & B Academic Interest H & B Need Achievement H & B Perception of Thinking Ability Stony Brook, Absence of Behaviour Stony Brook, Responds to Questions Stony Brook, Solicitation  E.S. 0.2545 0.4291 -0.5193 -0.5481 -0.0895 -0.5808 -0.3339 -0.2946 -0.4081 -1.2718 -0.1837 1.3813 1.3571 130  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 1.3494 0.6456 -0.8403 0.4472 1.5319 -0.6738 0.3576 -1.6279 0.5893 0.3320 -1.6587 0.5625 1.0015 0.8728 -1.1805 0.5322 -1.6939 0.3078 0.7640 0.5486 -1.4318 0.5962 0.8017 -1.3909 0.4654 0.6936 -1.5100 -0.0768 0.0366* -2.4668 0.7566 -1.3337 0.9663 2.6551 0.0166* 0.1076 2.6268 0.0358* 0.0358  Woodcock-Johnson Reasoning Woodcock-Johnson Math Cluster Canadian TBS - Total Math Woodcock-Johnson Reading Cluster Canadian TBS - Reading Vocabulary Canadian TBS - Reading Comprehension  0.2465 -1.1799 -0.4780 -0.6405 -0.4569 -0.5200  0.6600 0.0500* 0.3954 0.2628 0.4180 0.3576  -0.8481 -2.3609 -1.5838 -1.7583 -1.5614 -1.6285  1.3410 0.0011 0.6278 0.4774 0.6476 0.5886  Lower d scores are best below H & B Achievement Anxiety Stony Brook - Interference Stony Brook-Off Task Stony Brook - Minor Motor Movement  -0.5993 0.4824 0.7171 0.0896  0.2938 0.4180 0.2340 0.8808  -1.7138 -0.6813 -0.4659 -1.0586  0.5152 1.6462 1.9000 1.2378  Sarot, G.F. (1985). The effects of Feuerstein's Instrumental Enrichment on the academic achievement of educationally exceptional students. Unpublished doctoral dissertation, Northern Arizona University, A Z .  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American - Anglo and Mexican Mixed Ages reported as: Under 16.5 and Over 16.5 9-12 Low Special education students ( R S P & S D C ) M a i n l y male 1 year Fourteen instruments Pre-Post, Treatment Control, Repeated Measures design. Post-test and follow-up in year 2 Not classroom teacher Journal writing program  The study used a repeated measures design. Standard deviations were not reported so the effect size was calculated from an adjusted F-statistic. Treatment and control groups were assigned through the district's normal scheduling procedures. The control group received a journal-writing program. From the population o f special education teachers who volunteered, two were randomly selected for the treatment classes and two for the control classes. Information regarding teacher training was not presented.  Dependent Measures: >  Peabody Individual Achievement Test (PIAT) - Reading Comprehension, Reading Recognition and Mathematics  131  Effect Sizes Test PIAT Reading Comprehension  E.S. 0.2606  PIAT Reading Recognition 0.2183 PIAT Math 0.2489  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported.  Schneider, Sari D . (1996). Instrumental Enrichment and locus of control, self-esteem and academic performance in grade five. Unpublished master's thesis, University o f British Columbia, Vancouver, British Columbia, Canada. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  Canadian mixed - Asian Low Ten Five Mixed Mixed Mixed 4.5 months - 15 hours 1. Organization o f Dots Pre-Post, Treatment Control Not classroom teacher Regular curriculum  Inner-city schools with existing IE programs were invited to participate as the treatment group and matching schools without IE were invited to participate as the control group. T w o schools participated, one with IE and one without. The treatment group consisted o f fourteen students from two classrooms in the school offering IE, while the control group consisted o f 23 students from three classrooms in the school that did not. The groups were similar on demographic characteristics. There were two classroom teachers for the control group and for the treatment group in addition to the IE group. Good tests. Dependent Measures: > Coopersmith Self-Esteem Questionnaire (Coopersmith, 1975/1991) > Intellectual Achievement Responsibility Scale (Crandall, Katkovsky & Crandall, 1965) o Teacher Student Behaviour Checklist - Not used, as it was locally developed by the researcher. Effect Sizes Test Intellectual Achievement Responsibility Coopersmith  E.S. 0.3924 -0.2751 132  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.2502 -0.2780 1.0627 -0.9425 0.3922 0.4180  Shayer, M . , & Beasley, F. (1987). Does Instrumental Enrichment work? British Educational Research Journal, 13, 101-119.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  British Mixed 12 - 14 at the beginning o f the study Secondary Low Mixed Mixed 20 months (2 years) - 3 lessons per week - 150 teaching units Not reported Pre-Post, Treatment Control, Random selection Not classroom teacher Cognitive training called "Think Tank"  Shayer received training in IE and the L P A D from Feuerstein. He then had experience teaching a group o f students with severe learning disabilities. O n the basis o f this experience he provided initial training and regular inservice guidance o f the one IE teacher i n the study. Pupils i n a single class were assigned randomly to either the experimental or to the control group. Treatment and Control groups were independent. The control group received a cognitive training program called "Think Tank." Good tests.  Dependent Measures: > > >  > > o o  Piagetian battery (Fluid intelligence) Raven's Matrices (Fluid intelligence) Thurstone Primary Mental Abilities ( P M A ) test - Spatial Relations, Verbal (words), Reasoning (words), Perceptual Speed, Numbers (Number Facility). Numbers was calculated as part o f the Math achievement analysis rather than as part o f the A b i l i t y analysis. o Verbal (p), Reasoning (p) - Not used because I could not ascertain to what (p) referred. o Numbers - Not used because I was could not ascertain i f this was a measure o f the ability to reason with numbers or o f arithmetic skill. Neale Reading Test - Comprehension, Accuracy, and Rate N F E R Maths Attainment Richmond Basic Skills - M a p Reading, Graphs and Tables - Not used because means and standard deviations were not reported, Learning Potential Assessment Device ( L P A D ) (Feuerstein, Rand & Hoffman, 1979, 1980) Set Variations II - Not used as this measured pre-post changes o f observed deficient cognitive functions according to Feuerstein's theory.  133  Effect Sizes Test Raven's Standard Matrices Piagetian Battery Thurstone's P M A - Spatial Relations Thurstone's P M A - Perceptual Speed Thurstone's P M A - Reasoning (words) Thurstone's P M A - Verbal (words) NFER Maths Attainment P M A - Number Facility Neale Reading Comprehension Neale Reading Accuracy Neale Reading Rate  E.S. 0.1223 0.2213 0.5740 0.1343 0.2242 -0.1836 0.2428 -0.1416 -0.3838 -0.1517 0.6138  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.8336 0.7040 0.3320 0.8180 0.6966 0.7490 0.6744 0.8104 0.4902 0.7948 0.2984  -1.0104 -0.9138 -0.5806 -0.9986 -0.9110 -1.3176 -0.8930 -1.2746 -1.5258 -1.2849 -0.5441  1.2550 1.3564 1.7287 1.2672 1.3593 0.9504 1.3786 0.991557 0.7582 0.9815 1.7718  Skuy, M . , Lomofsky, L . , & Fridjhon, P. (1993). Effectiveness o f Instrumental Enrichment for pre-service teachers in a disadvantaged South African community.  International Journal of Cognitive Education & Mediated Learning, 3, 92-108  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design:  Instructor: Control Group:  "Coloured" South African English and Afrikaans speaking Mixed Average age: 18 First year students at a teacher training college. Mixed? Disadvantaged because o f segregated education system. Mixed 31 hours over sixteen weeks. 1. Organization o f Dots, 2. Orientation in Space I, 3. Comparisons, 4. Analytic Perception, 5. Categorization, 9. Family Relations, 10. Illustrations Pre-Post, Treatment Control Six o f 16 lecturers who had volunteered. College dramatic program.  The study used a quasi-experimental, two-group repeated measures design. The control group was chosen because the students had been selected for a college dramatic program and would not be available to be in the treatment group. The treatment group consisted o f the balance o f the first year students. Groups were "broadly the same ... i n terms o f age, S E S , intellectual and academic ability, gender and home language." IE teachers were six o f the 16 college lecturers who had volunteered to undergo an intensive seventeen-day workshop to become accredited IE teachers. Tests were good ones.  134  Dependent Measures: > Similarities: Combined Similarities subtests from the W I S C - R and the W A I S - R . > The Organizer (Feuerstein, Rand, Haywood, Hoffman and Jensen, 1983). It is part o f the Learning Potential Assessment Device ( L P A D ) . > Raven Standard Progressive Matrices ( R S P M , Raven, 1958). o Reading cloze test - Not used because it was a non-standardized, informal measure, o Teacher Attitude Questionnaire - Not used because it consisted o f open-ended questions. Effect Sizes Test LPAD Organizer Raven's Standard Matrices Similarities  E.S. 0.3057 0.1540 0.6670  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus -0.1032 0.7145 0.1416 0.4592 . -0.2535 0.5615 0.0012** 1.0825 0.2515  Socket, B . C . (1989). The effects of Instrumental Enrichment on cognitive performance, achievement, self-concept and maladjustment of adolescent females in residential treatment. Unpublished doctoral dissertation, Temple University, Philadelphia, PA. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Mixed 14-20 Secondary (Residential treatment) M i x e d ability, low achievement Socially and emotionally disturbed (EH) Female 1 year - 3 times a week for 45 minutes each period approximately 80 hours 1. Organization of Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception Pre-Post, Treatment Control, Random selection o f Ss Classroom teacher Structured games and activities  Subjects were randomly selected to be in either the treatment or control group. The control group received structured games and activities. The two teachers from the experimental group were selected by the school principal. Both teachers as well as the researcher received 5 days of training i n the rationale and implementation o f IE. Good tests. Dependent Measures: > Piers-Harris Children's Self-Concept Scale (Piers, 1984) - Total score > Raven's Standard Progressive Matrices (1977?). F test was an A N C O V A > California Achievement Test ( C A T ) ( C T B / M c G r a w - H i l l , 1986). F test was an ANCOVA 135  o o  Bristol Social Adjustment Guides (1970) - Under-reacting - N o t used. Total score not reported Bristol Social Adjustment Guides (1970) - Over-reacting - N o t used. Total score not reported Effect Sizes Confidence  Test Piers Harris Children's Self-Concept Scale Raven's Standard Progressive Matrices  Confidence  E.S. 0.2165 0.3093  CAT - Total Reading 0.5235 CAT - Total Language 0.6372 CAT - Total Math 0.4066  0.5686 -0.5264 Unable to calculate as only reported. Unable to calculate as only reported. Unable to calculate as only reported. Unable to calculate as only reported.  0.9595 F-statistics were F-statistics were F-statistics were F-statistics were  Stavros, D . (1989). Evaluation of the Instrumental Enrichment project, 1988-1989. Evaluation and Testing Department, Office o f Instructional Improvement, Detroit Public Schools. Detroit, I L . ( E R I C Document Reproduction Service N o . E D 361 462. Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Not reported Not reported - approximately 1 2 - 1 4 Thirty classrooms i n Middle Schools Low Educable Mentally Impaired ( E M I ) , Learning Disabled (LD) Not reported - mixed 1 year, three to five 50 minute periods per week Not reported Pre-Post, Treatment Control Not classroom teacher Regular curriculum  Experimental and control classrooms were matched. Teachers and staff were trained. This is a study that began in 1986 and was to continue in 1990. Dependent Measures: > Stanford Diagnostic Reading and Mathematics subtests, Form A , R e d Level o Cognitive Abilities Test, Forms D and E (1986). N o t used because only pretest scores were reported for this time period, o Absences from school, number o f students mainstreamed, positive changes i n teacher attitudes and feelings towards IE instructional practice, teachers' ratings o f changes i n 136  student behaviour, and parents' report o f positive changes in students' behaviour at home. Not used as these were informal, locally developed measures. Effect Sizes Test Stanford Diagnostic Reading  E.S. 0.159223  Stanford Diagnostic Math 0.603846  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus Unable to calculate as only F-statistics were reported. Unable to calculate as only F-statistics were reported.  Tillman, M . M . (1986). The effectiveness of Feuerstein's Instrumental Enrichment program in teaching thinking skills to learning disabled adolescents. Unpublished doctoral dissertation, Georgia State University, Atlanta, G A . Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American, mixed Middle to upper-middle class 11-14 6 - 8 . Private middle school Average to above average IQ. Academic achievement low in one or more areas. M i l d to moderate Learning Disabilities M i x e d , mainly male 1 year - 52 hours. 45 minutes per week. 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 6. Instructions Pre-Post, Treatment Control Not classroom teacher Current Events  The study used a repeated measures pre-post design. Experimental group obtained by randomly selecting one-half o f the students from each o f three middle school classes. The control group received current event activities. Treatment and Control groups were independent. Teachers were trained in the use o f the materials. The experimenter was one o f these teachers. Dependent Measures: > Standard Progressive Matrices (Raven, 1958) > Cognitive Abilities Test ( C A T ) , Multi-Level Edition, Form 3. (Thorndike & Hagan, 1983) - Verbal, Quantitative and Nonverbal batteries. (Grade 6: Level D , Grade 7: Level E , Grade 8: Level F). > Iowa Tests o f Basic Skills, Form 5 (Hieronymus & Lindquist, 1971)  137  Effect Sizes Test Raven's Standard Matrices Thorndike & Hagen, (CAT) Nonverbal Thorndike & Hagen, (CAT) Verbal Thorndike & Hagen, (CAT) Quantitative ITBS  E.S. 0.4166 0.6864 0.7245 0.4993 0.7238  2-tailed p 0.3222 0.1096 0.0930 0.2380 0.0930  Confidence Confidence Interval - Minus Interval - Plus 1.2436 -0.4103 1.5282 -0.1555 -0.1200 1.5690 -0.3315 1.3300 -0.1207 1.5682  Tzuriel, D . , & Alfassi, M . (1994). Cognitive and motivational modifiability as a function o f the instrumental enrichment (IE) program. Special Services in the Schools, 8, 91128.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used:  Design: Instructor: Control Group:  Israeli Low Average-12.3 Seven Mixed Mixed Mixed 2 years. 2 weekly hours; approximately 130 hours 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception, 6. Instructions, 8. Numerical Progressions, 9. Family Relations, 13. Orientation in Space II, 14. Classifications, 15. Organization o f Dots II, Cartoons Pre-Post, Treatment Control Not classroom teacher Regular curriculum  The IE program group was i n a school which had implemented IE for several years whereas the nonprogram group was in a school which had not introduced the IE program. Treatment and Control groups were independent.  Dependent Measures: >  o o  Learning Potential Assessment Device ( L P A D ) (Feuerstein, Rand & Hoffman, 1979, 1980) > Set Variations II - Total score used > The Organizer Test - Gain score used > Representational Stencil Design Test Raven's Progressive Matrices (Raven, 1959) - N o t used. Administered as a preintervention dynamic assessment only. The Intrinsic versus Extrinsic Orientation Scale (Harter, 1980, 1981) - N o t used as means and standard deviations were not reported and the analysis was a M A N O V A . The Perception o f Competence Scale (Harter, 1979) - N o t used as means and standard deviations were not reported and the analysis was a M A N O V A . 138  Effect Sizes Test L P A D Organizer (Gain score) L P A D RSTD L P A D Set Variations II (Total score)  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus  E.S. 0.5547 -0.0949 0.0021  0.0002** 0.5092 0.9920  0.2656 -0.3788 -0.2816  0.8438 0.1890 0.2859  Uditsky, B . (1984). Feuerstein's Instrumental Enrichment program, vocational skills, and adolescents with a mild developmental delay. Unpublished master's thesis, University o f Alberta, Edmonton, Alberta, Canada.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  Canadian, M i x e d Mixed 11-13 First year o f a secondary school for M M H Low M i l d l y developmentally delayed ( M M H ) Mixed 1 year - four to five times a week for 1 hour each; approximately 162 hours Organization o f Dots, Orientation in Space I Pre-Post, Treatment Control Not classroom teacher Regular curriculum  Students were randomly selected from first year classes and, at the beginning o f the school year, the school administration identified experimental and control classes. IE classes were taught by teachers trained in approved F I E summer workshops. One teacher had completed the first level o f training and the second had had more experience and had completed the second level o f training.  Dependent Measures: >  Wide Range Achievement Test (Jastak & Jastak, 1978) - The oral reading score was not used. Math and Spelling were used. > Schonell Reading Test - Used for the reading score o Vocational Task: Error Detection, Descriptive Discrimination, Circuit Board Construction, Overt Strategies - not used because they were locally developed o Vocational Evaluation Rating Scale - not used as it was a locally developed, informal assessment  Effect Sizes Test W R A T - Spelling W R A T - Math Schonell Reading  E.S. 0.1974 0.3692 -0.2605 139  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.6312 0.3844 0.5418  -0.6160 -0.4621 -1.1013  1.0109 1.2005 0.5802  Wilson, D . L . (1994). The effects of Feuerstein's Instrumental Enrichment on at risk seventh grade students. Unpublished doctoral dissertation, University o f Oklahoma, Norman.  Ethnicity: Socio-economic Status: Age: Grade level: Ability/Achievement Level: Population: Gender: Length of Treatment: FIE Instruments Used: Design: Instructor: Control Group:  American mixed (Indian, Afro-American, Caucasian) Mixed Approximately 13 Seven Mixed M i x e d - Special Education, Remedial Reading, Gifted. A t risk because o f poor behaviour Mixed Four months - 1 semester - five days a week for 55 minutes a day - approximately 73 hours 1. Organization o f Dots, 2. Orientation i n Space I, 3. Comparisons, 4. Analytic Perception Pre-Post, Treatment Control Not classroom teacher Regular curriculum  The study used a quasi-experimental pre-post control-group design with matched pairs. Students were matched according to discipline referrals, race, gender, special education category, grade point average, ability index scores (SAI) and I T B X scores. Pretest measures indicated that the two groups were comparable. N o information was given regarding teacher training; however, no one can teach IE without training.  Dependent Measures: > >  > >  Developmental Test o f Visual Motor Integration ( V M I ) (Beery, 1989) Self-Perception Profile for Children (Harter, 1985) - Six subscales: scholastic competence, athletic competence, social acceptance, physical appearance, behavioural conduct, global self-worth Iowa Test o f Basic Skills (ITBS) - Composite score Woodcock-Johnson Test o f Achievement - Dictation subtest o Absences - Not used as this was an informal measure o Discipline Referrals - Not used as this was an informal measure  140  Effect Sizes Test Developmental Test of Visual Motor In. ITBS - Composite Harter - Scholastic Competence Harter - Athletic Competence Harter - Behavioural Conduct Harter - Physical Appearance Harter - Social Acceptance Harter - Global Self-Worth Woodcock-Johnson Dictation  E.S. 0.8285 -0.0645 0.3640 0.2985 0.8118 0.3109 -1.1538 0.4735 0.0913  **p<0.01 *p<0.05 Reported Confidence Intervals are 0.95  141  Confidence Confidence 2-tailed p Interval - Minus Interval - Plus 0.0142** 1.4911 0.1659 0.6842 0.8650 -0.8131 0.2670 -0.2780 1.0060 0.3628 -0.3418 0.9388 0.0164 0.1494 1.4742 0.3422 -0.3297 0.9515 -0.4662 0.0012** -1.8414 -0.1721 1.1191 0.1498 Unable to calculate as only F-statistics were reported.  Appendix B Summary of Studies Not Used  Twenty excluded studies are listed below. Four were excluded because they did not have a control group; two compared two different types o f treatment to see which was better, but did not have a control group; 11 reported data in a manner that made analyses impossible; one used informal, non-standardized measures; one study included no data; and one study was available only on unreadable microfiche. Chinn, P . A . (1995). The long-term effects of Feuerstein's Instrumental Enrichment on the reasoning, intelligence, high school graduation rate, and post-school adjustment of 12 to 15 year old students with educable mentally handicapping and learning disabled diagnoses. Unpublished doctoral dissertation, University o f Kansas, Lawrence, K S o Subjects for this study were the same as used by Kenney (1984) i n his original investigation. The study was not used because the exact number o f students re-tested was not reported and statistical information was incomplete. It was not possible, therefore, to calculate effect sizes. Dufner, H . A . (1988). Effects of training in problem solving on the problem-solving abilities of giftedfourth graders: A comparison of the Future Problem Solving and Instrumental Enrichment programs. Unpublished doctoral dissertation, Texas A & M University, College Station, T X . o The study compared F I E and the Future Problem Solving programs. There was no control group. Bachor, D . (1988). D o mentally handicapped adults transfer cognitive skills from the Instrumental Enrichment classroom to other situations or settings? The Mental Retardation & Learning Disability Bulletin, 16, 14-28. o It was not possible to calculate effect sizes because the author presented data i n the form o f median ranks, and his analyses were M a n n Whitney U and W i l c o x o n Ranked Sum W . Howie, D . , Richards, R., & Pirihi, H . (1993). Teaching thinking skills to M a o r i adolescents. International Journal of Cognitive Education & Mediated Learning, 3, 70-91. o Data presented in this study was not suitable for calculation in a meta-analysis. This was a single subject research design. R a w scores were reported and graphed.  142  Kaniel, S., Reichenberg., R. (1992). Instrumental Enrichment - effects o f generalization and durability with talented adolescents. Gifted Education International, 8, 128-135. o The authors analyzed the data with two- and three-way A N C O V A , and although they reported means and standard deviations, they did not report the number o f subjects in each group. It was, therefore, not possible to calculate effect sizes. Livingstone, K . S . (1988). An investigation of the effectiveness of Feuerstein's Instrumental Enrichment program on field dependent seventh grade science students. Unpublished doctoral dissertation, University o f Wyoming, W Y . o Not used because the fiche was unreadable, and because there were two treatment groups: F I E and Learning Cycle training. There was no control group. M a x c y , P. (1990). Combining a comprehensive thinking program, Feuerstein's Instrumental Enrichment, with computer thinking programs. International Journal of Cognitive Education & Mediated Learning, 1, 45-61. o Data presented in this study was not suitable for calculation i n a meta-analysis. Results consisted o f counts o f correct and incorrect student responses. Nakra, O. (1990). The teacher as mediator: A n evaluation o f Feuerstein's Instrumental Enrichment program i n the classroom. International Journal of Cognitive Education & Mediated Learning, 3, 221-227. o In this study videotaped teacher behaviours for IE and non-IE teachers were analyzed. N o data were presented on how these teachers influenced the performance o f students on standardized measures o f ability, achievement, behaviour or self-concept. Narrol, PL, Silverman, FL, & Waksman, M . (1982). Developing cognitive potential i n vocational high school students. Journal of Educational Research, 76. 107-112. o The authors reported means, standard deviations and p-values for five classrooms, but not the number o f students in each group. It was, therefore, not possible to calculate effect sizes. Niles, R. (1989). Effects of adult modeling and adult direction on impulsive behavior of learning disabled adolescents. Unpublished doctoral dissertation, Fordham University, N e w York. o The study compared F I E and the Meichenbaum adult modeling approaches. There was no control group. Offenberg, R . M . (1992). A study of the effects of Instrumental Enrichment on middlegrade, minority students. Report No. 9225. Philadelphia, P A . Philadelphia School District ( E R I C Document Reproduction Service N o . E D 361 462. o There was no control group; instead the amounts o f F I E instruction varied. In addition the data were evaluated with regression analyses and results were reported as betaweights.  143  Pachal, Judith. (1986). The use of Instrumental Enrichment with hearing impaired adolescents. Unpublished master's thesis, University o f Alberta, Edmonton, Alberta, Canada, o This is a qualitative study o f problem-solving deficiencies i n four hearing-impaired adolescents. Rothen, W . (1989). Interim evaluation of the Instrumental Enrichment (I.E.) Program. Unpublished report, Vancouver, B C . o This was a pre-post study with no control group. R u i z Bolivar, C . (1983). The effects of the Instrument Enrichment (sic) Program on cognitive and non-cognitive factors on the subjects of different socio-economic levels. Ciudad Guayana, Venezuela: Office o f Social and Cultural Development. o The section presenting data were omitted in the copy I was able to obtain. Silverman, H . , & Waksman, M . (1988). Feuerstein's Instrumental Enrichment: Elicitation o f cognitive interaction in the classroom. Canadian Journal of Special Education, 4, 133-149. o This is a study o f the effect o f using IE as a means o f improving the questioning style and general teaching style o f teachers. Attempts were made to administer standardized measures o f reading and mathematics to students to analyze how the quality o f teacher questioning effected their achievement. The students objected to the measures, however, and attempts to formally measure academic achievement were abandoned. Skuy, M . , Mentis, Martene, Durbach, F., Cockcroft, K . , Fridjhon, P., & Mentis, Mandia. Crosscultural comparison o f effects o f Instrumental Enrichment on children i n a South African mining town. School Psychology International, 16, 265-282. o There was no specific control group i n this study. Instead there were comparisons among groups o f African, "Coloured," white English-speaking, and white Afrikaansspeaking students. Skuy, M . , Mentis, Mandia, Arnott, A . , & N k w e , I. (1990). Combining Instrumental Enrichment and creativity/socioemotional development for disadvantaged gifted adolescents i n Soweto: Part 2. International Journal of Cognitive Education & Mediated Learning, 2, 93102. o The study was analyzed by means o f an A N C O V A with three groups, and although mean scores for the IE and control groups were reported, the number o f subjects i n those groups was not. It was, therefore, not possible to calculate the effect sizes. Stavros, D . (1987). Evaluation of the Instrumental Enrichment project, 1986-1987. Evaluation and Testing Department, Office o f Instructional Improvement, Detroit Public Schools. Detroit, I L . o This paper did not give data that could be used in this analysis. Assessments consisted o f locally developed measures o f absences from school, number o f students mainstreamed into regular classes, teachers' positive attitudes towards students and parents' reports o f positive changes in the students. There were standardized measures o f reading and mathematics; however, only pretest results were reported. Results o f the 1988-1989 year are presented above i n the Summary o f Studies Used.  144  Thickpenny, J.P., & Howie, D . R . (1990). Teaching thinking skills to deaf adolescents: The implementation and evaluation o f Instrumental Enrichment. International Journal of Cognitive Education & Mediated Learning, 3, 193-209. o The study consisted o f a single subject research design plus a t-test between pre- and post-measures o f both experimental and control groups. The two groups were not directly compared at the post stage. Yitzhak, V . (1981). The effect of Feuerstein's Instrumental Enrichment program on the cognitive reasoning of retarded performers as measured by Piaget's conservation tasks. Unpublished doctoral dissertation, University o f Toronto, o N o t used because the microfiche was unreadable.  145  Appendix C  Results from Random Effects Analyses In the following table, the two-tailed p-value gives the probability that the effect size is not zero. In addition, i f zero is within the 95% confidence interval the effect size does not significantly differ from zero. Note that the two-year and follow-up results from Samuels (1984) are not included with the larger group. This is because the same subjects that were in her 1-year measurement were used in the second- and third-year measurements, and the results, therefore, would not be independent. This rationale also applies to second-year and follow-up measurements by Church (1994), Perry (1986), and Rand, Mintzker, et al. (1981). (See Appendix A . )  146  £1 10  ON  oo  m o  0  ©  NO  cn  m rs O  as i n  3 rs  «o  is  00  ho  o oo o  cn  ON NO 00  in  °  io  ON  cn  ON  ©  |ON|•t  ON 00 ON  ON  ON  ON  S  >r, rs  ON  ON  I ON o| S  e  eu  ctt  >  >  **  # *  '3 O o  '3 o o  rs o ©  H o  H o rs o  N l  00 CO ON  o rs m o  w m  o o  <  in r--  * '  * NO rs o o  •a 00 o o  CO  CO i n  rs  tr-  ee,  rs  ©  I  ca -*-»  c/)  'S D 3 ti.  o X ca  2 ca  CO  s  CO  ca , <L> ca <£ >, Irs ca rs I >- , ca <U >i' 3  °5 ca O e ca >,CO rs .T3 C  ca  T3 C  ca Pi  ^ 5  >  o—  JC CO  — <a  >J  rs u o 3 t/5 S ca u co w = * "ca ffi I  u  c  ca e <u c PQ A>  OS  c c  S i ?  PQ  s-  4)  m I _>> D c/i ca <u PQ  >, 3 q ^5  ca  to  s  ca o -w  CO  tf  c o  Ja  3  ca  OH  o  s ca  eS  a £ a a a  -<  "J3 -a I  •a  o Z  147  .5 g> 3 .S5E 2 3 m PQ pi I  PH  s  Vi  IS li c" l-c '5 V3 ca 3 b rs +H  in  o  o m o o  o  ft  Z,  o  I  o  o ON  in CN  o  o CD  s  o Cfl  S  !. o u  cd  cc  co  cS  >. CN  s  o  c  o  >  S3  tu 3 N  s-  o  <  CN  a  J5  "O  ©  1 5T  u  H  o I s CO  « w «<  3  cd  00  CH*  7& 5. tu c  J  rj  1=1  CU  148  co  CD 3 U U,  _C  o 3  o  o  CJ  -g la cj  lu CN  ra  o  I o\ ce a  NO  uo  0©  ,CJ  cn  C  ca  P-  g  "  "O L3  b ^ IH  U  £ K ca E >- 5 c£2 3 CN  CN CN O  cn  o  c3  rrl  i'l  NO  NO  CJ  _ CJ  c -5  C  J2 C+H  E o  cj  X-  O "O  a 3 •s S  * CN  O  c ca  cn cn  00  E  o  CTN  H  ^  CD CD OO  a  o o  a  CJ  00  2  v.  ^ e  §1  CN  O <C  <  T3  cj  h =el's. 2 « ^  ea  s ^»  ca JM c o  eg  I  c/.  CC >—•  *i cj  DS  Lc u  a O)  cs  ,  C "ea CJ C *" o  1-  u  00 O H 3  ^  1^ SH  c/i  <U  1-=;  .a .S  u ~  1 3  cn  ea ca p ,  ca  3  ca ,  CN  CJ  c/i CJ  cn B  E  0>  ca  ^  C/5  o rn  S  H3 fe LH S-  u  .a l DC H  w  1— ^  -C  cn  1  CD  53 ^ CD  <u  ^  CQ  3 -a  X H  0)  01  C«  O <D cn f—  OJ  ., CN  a  ca  ca  55 ca 3.  CJ  in  S SHc°  ca  CS  C3  rn  -c in  •3  8  cs s-  H ca 0)  CQ CQ  149  C*M CO  a  cs  H 1. cs  fa  ll 0  c  Q> O  Ui .1— «  CJ  5  S  cj  cn  _ —  cj to CO  lu  I ' 00 ON o ,r  NO 00  ro CN  o  in 00 ON o  o I ONo a NO B  NO m N O ©  ON C N OA ro s ©  © 00  ON  ©  s  SB  <o ©  ON © s NO  I ONC N S 8  ©  li CJ  s  "3  >  o oo  TT  _ej  H  CO  o CN ON  ©  CN CN ©  ©  "3  NO OO  hs  CN  , '3 T© f  H  ©  o o o  CO  Tf © CN  co  r-  H  CN  ON o o  CN  NO  CO ©  ro NO  cj CJ  a «  CN  o  -3-  I c  te Ol  ej  CD T3  "« e  J=  on "c '5  CO  o  GO S.  o co  « a ©  CJ  CS  -a c cS ^  "3  e cj  a  ^ j= CO  CJ  o  CN  £ = cS o  CJ CJ  a CJ  al  E  c j= o co  «S &  JC CS CS u,  cs  '3  i~  <u  a  i-  m  CJ CJ  I  < > t  PH  <  E in  3  cj  u s  a a  c o  3 +H  s  CJ  JC  o CJ  «  O -4-*  CJ  CO  u S  o O  "3  >>  ^ej  •§  -o c  J=  o  CS  ai tf  o  c  eS 4=  o U  CJ CJ  o  a  3  CJ  •9  «a  3  o  CS  o  §  IS  s2 «i 150  U  CJ  co  "a a o PH  -a c cS — cS O  PH  01  T3 cs ej  C o  •o a cs  3  cS cDS  co  c o  CJ CJ  a  cs  a CJ CJ CJ  CJ  o  CO  C o  HH  u \m\ in ON  IT) ON  NO  co  o  o in m ON  o  co  cn S PH  in CN in  ^ H  Tf  IT) ON  00  CO  CN Tmf C N ON rC N ' 1 r© d  cn © S Tf ©  00  ©  HH  o  u ION  cn NO S co  in roo  ON ON  ON  ro  C fl S  Min  ON  ©  ON  CN ©  1  NO  CO  o  m  ON  cn S  CN d d• NO  Min  m  ID  I ON CO CN  I  in co CN r- © CN CO T f CN T CN f ©  © i  . o  §- 0)  © © • 1  2 S 6X) CD  *- E  = -5 o ca  JS  "3  >  >•  a  CN CN in o ON CN 'cs in T f © d d H cs  Ti-  led  O ON Tf  , '3 ©  ro  H  H  CN  S J CD cS  "«  CL  •o CN  SI  CU  eu  js  Tf  ©  NO  NO  '3 H  ON  ©  m in i-H  ©  © ©  o c  JS  ea  NO  d  CN  u >cn CD  CD CD JS JC  00 *=  '3  t» oNO oON  m  NO 0©  eo ro  Tf  Tf  o CO 0 0 d d  w  NO  W  ©  m m  NO  CD  T3 ca  ON Tf  © in © © © © © • •  cCDn  «  JS J3 Q _ CD 3 JS  CD  C  o O co c o CD  8 « g  •o s  CD CD-  JS  in  Tf  !_ o  ca  ca  x> c u c  "1  ct-<n  an  S "3 V  e <u  06  ler.  O  an  H  ca  <<  3  o oo  cn  S  E  , 1>  .SP  i—  I ca  CD  CD  CD  + H  ca  tf .23 c  l i  _9  ca ca CD CD >>  CN 1 CO1 cCDn CD 3  3  o E E ca ca oo oo  ca  3  E  ca  ss o  • mm  > cs  J=  eu  pa  c u  ss ss  ^ca CD  oo  cn  "CD 3  E E ca CD ca oo H- oo  151  J=  CN  s  i  *-<  CD  C ES  3  •*-»  s  o  1 PQ  CU CU  z  •  CD  cn  >-,  E c  CD  PU  CD  1  ca  oo  CD J S  >  ca so  JD  m ©  CD  © ©  > H  + H  >  3  CD  e8 >  <+-c T3  8 ®  Jj  c  CD CD J S JS J J SS >  .5  § E •r o ca  c  CD  J=  1 a 1 '33 1/3  "CD 3  ca  ye  cn eu  ca  CD CD JS  JS  CN o co J*  rry, Sa uel ea uerstei Pe ry, Sam els real muels ar aft  •o e cs  CD >>  ors  o  tf  els - 2 ye after trea  ca £ tf  CD J^ S  ca  uer  e  _o tf  "5 ^  ca ca  c C3 o IS  § sS C  oo oo <;  |  CD JS  Appendix D  Significant Results Information A Table with the following information from combined Random Effects analyses that had significant results: Authors, tests used, type o f control groups, single or grouped effect sizes with a statement regarding significance, and a statement regarding whether or not the study had been published.  There were significant results for the following:  Full Scale Ability after 1 or 2 years of treatment (d=Q.519i *,p<.01). Author Tests Used Effect Size & Novel Control Significant?  Published?  Feuerstein Funk Kenney Maxcy Muttart Pasquier Perry Samuels  Primary Mental Abilities Test Total Otis-Lennon Woodcock-Johnson Reasoning Test of Cognitive Skills Primary Mental Abilities Composite SPE 129 Otis-Lennon Woodcock-Johnson Reasoning  Yes  0.5307, Yes  Yes  No No No No  0.9657, Yes 0.6143, Yes 0.3734, No 0.8946, No (Almost) 1.5121, Yes 0.0830, No -0.1444, No  No No No Yes  No No No  Verbal Ability after 4 weeks to 2 years of treatment (d=1.4050, p<.01). Tests Used Novel Effect Size & Significant? Control? Eichler Watson-Glasser No 0.9668, Yes Hall Test of Social Inference No 0.1627, No Samuels -0.1804, No WISC-R Verbal No Shayer & Thurstone's PMA - Reasoning & Yes 0.0200, No Beasley Verbal (words) Skuy et al. WISC-R & WAIS-R Similarities Yes 0.6670, Yes Tillman Thorndike & Hagen (CAT) Yes 0.7245, No Verbal  152  Yes No No  Published? No No No Yes Yes No  Non-Verbal A bility after 3 months to 2 years o f treatment (^=0.2414, p<.01). Tests Used Novel Effect Size & Published? Control? Significant? Brainin Thorndike & Hagen (CAT) No 0.3221, No No Nonverbal Feuerstein Embedded Figures, Human Yes 0.3913, Yes Yes Figure Drawing, Lahy, Terman (Just) Concept Formation - nonverbal, D-48, Porteus Maze 0.3542, No Hall Ravens P . M . , Standard No No 0.2514, Yes Kenney Ravens P.M., Standard No No Maxcy Ravens P.M., Standard No 0.3039, No No Robinson -0.2199, No No Ravens P.M., Standard No Rusling No -0. 3200, No No Ravens P . M . , Coloured and Standard Samuels WISC-R Performance Scale No -0.2797, No No Shayer & B Yes Ravens & Piagetian Battery Yes 0.1716, N o 0.1540, Yes Yes Skuy et al. Ravens P.M., Standard Yes Tillman Ravens P . M . , Standard & Yes 0.5486, No No Thorndike & Hagen (CAT) (Almost) Nonverbal  Visual Perception after 5 months to 2 years o f treatment (cH).4237, p<.05) Novel Effect Size & Tests Used Control? Significant? Feuerstein Kuhlman-Finch Postures, Yes 0.7014, Yes modified Hall Matching Familiar Figures No 0.6477, Yes. Rusling 0.2808, No Motor-Free Visual Perception No Shayer & Thursone's P M A - Spatial Yes 0.3481, No Beasley Relations & Perceptual Speed  Visual-Motor Ability after A a year o f treatment (*/=0.7046, p<.01). Novel Effect Size & Tests Used Significant? Control? Rusling No 0.6085, No Porteus Maze (Almost) Wilson Developmental Test of VisualNo 0.8285, Yes Motor Integration  Published? Yes No No Yes  l  Published? No No  Visual-Motor Ability 1 year after Church's yearlong treatment (d=1.6775, X . 0 1 ) . Effect Size & Published? Tests Used Novel Control? Significant? Church 1.6775, Yes No Bender Gestalt & Visual Aural No Digit Span  153  General Achievement after 4 months to 2 years o f treatment (g/—0.2571, p<.05).  Tests Used Church Feuerstein Hall Muttart  Iowa Test of Basic Skills Project Achievement Battery Geography, Nature, General Knowledge PIAT - General Information Canadian Test of Basic Skills  Tillman Uditsky Wilson  Iowa Test of Basic Skills WRAT - Spelling Iowa Test of Basic Skills  Novel Control? No Yes  Effect Size & Published? Significant? -0.0645, No No 0.1504, No Yes  No No  0.4759, Yes 0.9738, No (Almost) 0.7238, No 0.1974, No -0.0645, No  Yes No No  No Yes No No No  Math Achievement after 1 and 2 years o f treatment (fif=0.2914, p<.05).  Tests Used Funk Hirsch Mulcahy Samuels Shayer & Beasley Tillman Uditsky  SAT - Total Math WRAT & Comprehensive TBS Math Canadian Achievement Test Math Woodcock-Johnson & Canadian TBS Thurstone's PMA - Number Facility Thorndike & Hagen CAT Quantitative WRAT - Math  Novel Control? No No  Effect Size & Published? Significant? 0.9367, Yes No 0.4165, No No  No  0.0974, No  No  No  -0.5675,Yes  No  Yes  -0.1416, No  Yes  Yes  0.4993, No  No  No  0.3692, No  No  Near Transfer after 1 Vi years (d= 1.6123, P<.01).  Tests Used  Novel Control? No  Effect Size & Published? Significant? 1.6123, Yes No  Far Transfer after 1 A years (d=T.0767, p<.0 I). Tests Used Novel Control? Jensen French Kit & parts of IE No curriculum  Effect Size & Published? Significant? 1.0767, Yes No  Jensen  French Kit & parts of IE curriculum  l  154  Locus of Control - Intellectual Achievement 4 A months - 1 year (d=03318, p<.05). Published? Tests Used Novel Effect Size & Significant? Control? Kenney Intellectual Achievement No 0.4505, Yes No Responsibility Questionnaire No Maxcy Intellectual Achievement No 0.9215, No Responsibility Questionnaire (Almost) Mulcahy -0.1375, No No Intellectual Achievement No Responsibility Questionnaire Schneider 0.3924, No No Intellectual Achievement No Responsibility Questionnaire l  General Motivation, and Attitudes and Motivation for Learning 1 year after 2 years of treatment in Samuels' study (d=-0.4787, p<.05). Published? Novel Effect Size & Tests Used Control? Significant? No Samuels Holtzman & Brown Survey of No -0.4787, Yes Study Habits/Williams Perception of Thinking Abilities - seven subtests  155  Appendix E  Nonsignificant Results Information A Table with the following information from combined Random Effects analyses that had non-significant results: Authors, tests used, type o f control groups, single or grouped effect sizes with a statement regarding significance, and a statement regarding whether or not the study had been published.  There were non-significant results for the following:  Full Scale A bility 1 year following 2 years of treatment (d= -0.0520) Author Novel Effect Size & Tests Used Control? Significant? Perry Otis-Lennon School Ability No -0.0197, No Woodcock-Johnson Reasoning, No -0.2555, No Samuels WISC-R Full Scale  Published? No No  Negative Visual Perception 1 year after 5 months of treatment («i=0.1306). Author Tests Used Novel Effect Size & Published? Control? Significant? Rusling Benton Visual Retention-Error No 0.1306, No No scores; Trail Making Test - A & B-Time Scores m  LPAD 4 weeks, 1 year and 2 years of treatment (d= 0.4960) Author Effect Size & Tests Used Novel Control? Significant? Eichler 1.4552, No RSDT & Set Variations I No Skuy et al. Organizer - post-test scores Yes 0.3057, No Samuels Numerical Progressions No 0.7917, No Organizer - gain score, RSDT, No Tzuriel 0.1920, No Set Variations II - total score  LPAD 1 year after treatment (d= 0.3409) Author Tests Used Samuels  Numerical Progressions, Set Variations II  156  Novel Control? No  Published? No Yes No Yes  Effect Size & Published? Significant? 0.3409, No No  Reading 1 and 2 years of treatment (d= 0.0015) Author Novel Tests Used Control? Brainin No CTBS - total Feuerstein Yes PAB Funk SAT - Comprehension No Hirsch Mulcahy Perry Samuels Shayer & Beasley Uditsky  Brigance, Comprehension TBS, WRAT - word recognition Canadian Achievement Test PPVT Woodcock-Johnson cluster score, CTBS Comprehension & Vocabulary Neale Comprehension, Accuracy & Rate Schonell  Reading 2 years of treatment (d= 0.0299) Author Tests Used Perry Samuels  PPVT Woodcock-Johnson cluster score, CTBS Comprehension & Vocabulary  Reading 1 year after treatment (d= - 0.5380' Author Tests Used Samuels  Woodcock-Johnson cluster score, CTBS Comprehension & Vocabulary  Math 2 years of treatment (d= - 1.0409) Author Tests Used Samuels  Woodcock-Johnson cluster score, CTBS total score  157  Published? No Yes No  No  Effect Size & Significant? -0.3877, No 0.0177, No 0.5043, No (Almost) 0.0585, No  No No No  -0.0694, No 0.2524, No -0.2571  No No No  Yes  0.0171, No  Yes  No  -0.2605, No  No  Novel Control? No No  Effect Size & Significant? 0.0495, No -0.1252, No  Novel Control? No  Effect Size & Significant? -0.5380, No  Novel Control? No  Effect Size & Significant? -1.0409, No (Almost)  No  Published? No No  Published? No  Published? No  Math 1 year after treatment (d= - 0.7977)  Author Samuels  Tests Used  Novel Control? No  Woodcock-Johnson cluster score, CTBS total score  Effect Size & Published? Significant? -0.7977, No No (Almost)  Self-Esteem, Confidence in Personal Success 4 months to 2 years of treatment (d= 0.0079)  Author Maxcy Mulcahy Rand, T et al. Schneider Wilson  Coopersmith Coopersmith, Harter Levidal  Novel Control? No No Yes  Effect Size & Published? Significant? 0.2344, No No No -0.0358, No 0.0673, No Yes  Coopersmith Harter - Global Self-Worth  No No  -0.2751, No 0.4735, No  Tests Used  No • No  Academic Self-Confidence 4 months to 2 years o f treatment (d= 0.0700)  Author Mulcahy Muttart Rand, T et al. Wilson  Tests Used Coopersmith, Harter Brookover, St. John Scale, Achievement Self-Esteem Levidal Harter  Novel Control? No No  Effect Size & Published? Significant? 00928, No No Yes -0.0205, No  Yes  -0.0868, No  Yes  No  0.3640, No  No  Self-Confidence of Physical Competence and Appearance 1 year, and 4 months o f t r e a t m e n t ^ 0.1204)  Author Mulcahy Wilson  Tests Used  Novel Control? No No  Harter Harter  Effect Size & Published? Significant? 0.1041, No No 0.3048, No No  Social Acceptance, Skills and Popularity 1 year, and 4 months of treatment (d- 0.6306) Author Tests Used Effect Size & Published? Novel Significant? Control? Mulcahy Harter No 0.1109, Yes No Wilson Harter No -1.1538, Yes No  158  Self-Concept 1 year of treatment (d= 0.3371) Author Tests Used Hall Muttart Socket  Piers Harris Lipsett Piers Harris  Novel Control? No No Yes  Effect Size & Published? Significant? No -0.0323, No Yes 1.1733, Yes 0.2165, No No  General Motivation, Attitudes, and Motivation for Learning 4 weeks to 2 years of t r e a t m e n t s 0.1865) Author Tests Used Novel Effect Size & Published? Control? Significant? Eichler Picture Motivation, Haywood No 1.29, No No Maze Yes Rand, T. et Levidal Yes -0.0315, No al. Samuels Holtzman & Brown Survey of No -0.7141, Yes No Study Habits/Williams Perception of Thinking Abilities - seven subtests  General Motivation, and Attitudes and Motivation for Learning after 2 years of treatment in Samuels' study (d= -0.2401). Effect Size & Published? Tests Used Novel Control? Significant? Samuels Holtzman & Brown Survey of No -0.2401, No No Study Habits/Williams Perception of Thinking Abilities - seven subtests  Behaviour a ter 1 year of treatment (d- 0.1160). Tests Used Novel Control? Samuels Stony Brook Observation Code No - four subtests  Behaviour a ter 2 years of treatment (d= 0.3690). Tests Used Novel Control? Feuerstein Classroom Participation Scale - Yes two subtests Samuels Stony Brook Observation Code No - four subtests  159  Effect Size & Published? Significant? 0.1160, No No  Effect Size & Published? Significant? Yes 0.8893, Yes -0.1797, No  No  •  Behaviour 1 year after treatment (d= 0.8014). Tests Used Novel Control? Samuels Stony Brook Observation Code No - three subtests  Effect Size & Significant? 0.8014, No  after 1 year of treatment (d= - 0.0074). Tests Used Novel Effect Size & Significant? Control? Burks Behaviour Rating Scales No 0.1647, No - nineteen subtests Stony Brook Observation Code No -0.3034, No - f i v e subtests  Published? No  Negative Behaviour  m  Perry Samuels  Published? No No  jijj  Negative Behaviour Feuerstein Perry Samuels  after 2 years of treatment (d= - 0.0565). Tests Used Novel Effect Size & Control? Significant? Classroom Participation Scale - Yes -0.1789, Yes four subtests Burks Behaviour Rating Scales No -0.0280, No - nineteen subtests Stony Brook Observation Code No 0.1873, No - f i v e subtests  1 year after treatment (d- 0.1495). Novel Effect Size & Tests Used Significant? Control? Achievement Anxiety & Stony No 0.1495, No Brook Observation Code - three subtests  Published? Yes No No  Negative Behaviour  m  Samuels  Published? No  With measures of negative behaviour, the lower the effect size, the better the performance for the treatment group.  160  Appendix F  Effect Sizes from F-Tests, t-Tests and Reported Effect Sizes The following table gives single effect sizes calculated from F- and r-tests, and effect sizes where they alone were reported. Because the authors did not report descriptive statistics, it was not possible to combine them with the others i n the meta-analysis. The only effect sizes listed w i l l be those that Cohen (1988) considered to be medium sized (0.50) or above.  Verbal Ability Authors: Haywood, -1.5 years Test Primary Mental Abilities - Reasoning  E.S.'  2  1.2477"  Non-Verbal Ability Authors: Bruce & Martin, Haywood et al., Martin Test E.S. Raven's Progressive Matrices Raven's Standard Progressive Matrices Raven's Standard Progressive Matrices  0.5128* 0.6991* 0.8963  Visual Perception Authors: Arbitman-Smith et al., - 60 hours Test E.S. Hiskey-Nebraska, Spatial Reasoning  Visual-Motor Ability Authors: Church, - 1 year Test Bender Gestalt Visual Motor Integration Visual Aural Digit Span  0.7362*  E.S. 1.0581** 0.5440**  Effect Size * Significant at the 0.05 level ** Significant at the 0.01 level 1 2  161  Academic Achievement: Reading Authors: Socket, -1 year  Test |CAT - Total Reading  E.S. 0.5235  General Academic Achievement: Composite Scores, General Information, Language Authors: Haywood et al., Socket - 1 year, 1.5 years  Test Peabody IAT - Information Stanford Achievement - Language CAT - Total Language  E.S. 0.5150 0.5518 0.6372  Academic Achievement: Math Authors: Haywood et al., Hirsch, 1 year to 1.5 years  Test KeyMath - Math Reasoning Comprehensive TBS - Math  E.S. 0.5376 0.7426  Self-Esteem Authors: Jensen, 18 months Test Piers-Harris, Sample A Piers-Harris, Sample B  E.S. 0.6718* 1.3263**  Attitudes and Motivation Authors: Blagg, Jensen, McRainey -approximately 20 hours, 18 months, 2 years  Test E.S. 16PF - Teachers (Low = Humble, High = Assertive) 1.2270** 16PF - Teachers (Low = Tough-Minded, High = Tender-Minded) -1.0661** 16PF - Teachers (Low = Trusting, High = Suspicious) 0.9876* Haywood Maze Test, Sample C 0.5521* Classroom Environment Scale 0.6716* 13  14  Note that on the 16PF measures, the teachers scored i n such a way as to indicate that they were assertive, tough-minded and suspicious.  13  1 4  Jensen considered this test to be a measure of intrinsic motivation. Students assessed the social climate in the classroom.  162  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0055041/manifest

Comment

Related Items