An Overview of Problems Encountered in Aptitude-Treatment Interaction (ATI) Research for Instruction
KAREN ROSENKRANTZ SHAPIRO U. S. public education has attempted to equalize the opportunity for all students to develop to their potential. That it has failed is documented in such national evaluations as the Coleman Report (1966). Schools apparently provide few experiences to compensate for the unequal competencies with which students enter the system. Several strategies for the improvement of education have been based on the individualization of instruction. Students' characteristics focused upon for individualization vary according to instructional approach. In the mastery approach all students work through the same learning module but at their own pace. Slower students are afforded as much time as necessary to learn the material, and faster students avoid boredom by moving on to new modules. Increased individualization is also a characteristic of programed instruction (PI) and computer-administered instruction (CAI). In these approaches, the programs of students are guided by their understanding of the previous steps in inKaren Shapiro is a doctoral candidate in the Institute for Communication Research at Stanford University, California 94305. AVCR VOL. 23, NO. 2, SUMMER 1975
227
AV COMMUNICATION
R E V I E W 1975 S U M M E R
:
228
struction. Programs with multiple branches provide frequent "microadaptations" or individualizations of the instruction according to the current competencies of the learner. Aptitude-treatment interaction research can be regarded as a further extension of this trend toward individualization. In addition to speed of learning, past school experiences, and current comprehension, the researcher may consider personality traits, mental abilities, cognitive styles, or any other "characteristic of the person that forecasts his probability of success under a given treatment" (Cronbach & Snow, in press). ATI research in education asks the following questions (Snow, 1970): "Assuming that a certain set of outcomes from an educational program is desired and considering any particular instructional treatment, in what manner do the characteristics of learners affect the extent to which they attain the outcomes from each of the treatments that might be considered? Or, considering a particular learner, which treatment is best for him?" By extrapolation, ATI research presumably would produce a matrix of aptitudes by tasks with different treatments occupying the cells (Salomon, 1970). Such a matrix would enable teachers to locate a particular student by aptitude, choose the necessary learning task, and thereby determine an appropriate treatment. Although no one seriously envisions such matrices to be at the educator's disposal for all situations, proponents of ATI research do hope to provide practitioners with some "assistance in guiding their students. If this goal is to be realized, researchers must avoid the obstacles encountered in past research. In the following pages I enumerate the problems of ATI research and discuss several "real-world" factors which should be considered. Finally, I suggest a direction for future research which may prove more profitable. PROBLEMSOF PAST RESEARCH
Why have educators gained so little from ATI? 1. Research context. The majority of studies reported in ATI research are laboratory-based rather than field-based. This immediately limits their applicability to the school situation. Laboratory conditions are generally unattainable in the classroom. The precision and control of laboratory manipulations rarely translate into real instructional treatments. 2. Lack of theory. Minimal theory-based research has in-
AV OVERVIEW OF PROBLEMS IN ATI RESEARCH
:
229
tentionally probed for aptitude-treatment interactions. Often researchers looking for interactions administer a large battery of aptitude tests to subjects and examine correlations for each score and task performance in the hope of detecting significant interactions. Unfortunately, many of these results are spurious products of chance and would not stand the test of replication. Researchers also come upon interactions accidentally as they analyze their data and publish their findings as suggestions for future research. For the most part, interesting results which do emerge are not pursued, expanded upon, or tested in real-world situations. 3. Methodological weakness. Aptitude measures are frequently developed by individual researchers and are given no tests for reliability or content validity. Rarely do researchers attempt convergent and discriminant construct validation (Campbell & Fiske, 1959), nor do they attempt to align their constructs with those of other researchers. As a result of this lack of coordination, anyone hoping to use research results must have a sophisticated understanding of the literature. The significance of findings varies with the differences in operational definitions of terms employed in the study and these are not often made explicit. Thus, multiple investigations in areas such as programed instruction have led to inconsistent results. For example, in thirty-odd studies comparing programed instruction with conventional instruction, Cronbach and Snow (in press) found 14 studies showing no interaction, 12 studies found PI more beneficial for lowaptitude students, and 5 studies demonstrating that PI improved the performance of high-ability students. Practitioners can only be confused in their attempts to draw inferences from such work. 4. Inadequate data analysis and reporting. In reviewing past reports of ATI, Cronbach and Snow (in press) comment on the omission of regression analysis where it would have been the preferred methodology and bemoan the absence of data which would have allowed others to undertake such analysis. Experiments are frequently executed with small samples, thus increasing the difficulty of detecting an ATI and leading to inconclusive results. Cronbach and Snow suggest that samples of 100 subjects or more are necessary to test for interactions between aptitudes and treatments.
AV COMMUNICATION
R E V I E W 1975 S U M M E R
:
230
5. Diversity of researcher orientation. Researchers have turned to the ATI approach from several areas of inquiry. One group has focused on questions concerning particular aptitudes. These researchers measure subjects on the aptitude of interest and employ various treatments in order to validate the aptitude construct. Performance differences between subjects with high aptitude scores and subjects with low aptitude scores support the existence of the aptitude construct that predicted the differences. Aptitude-treatment interaction research also has evolved from the work of treatment developers. The introduction of media such as videotape, closed-circuit television, film, and computers into the educational environment has stimulated investigation into their most beneficial applications. Here again it is the fascination of the researcher which determines what will be studied; contact with a particular medium or treatment precedes consideration of task or aptitude. Until recently, the questions asked from this perspective were too broad to prove useful. Gross comparisons of live teaching with television instruction or programed instruction were unable to isolate consistent advantages across situations and contents (Jamison, Suppes, & Wells, 1973). More productive work from this orientation provides analysis of the specific attributes of the treatment media which will be used for task accomplishment (see Salomon, 1970, 1974). A third major impetus for aptitude-treatment interaction research has been the specification of educational goals or tasks. Educational researchers and practitioners have access to a multitude of treatments which profess to facilitate, if not accomplish, goal achievement. Focusing on a particular task, and given a limited variety of treatments, it is possible to search for the aptitudes that determine success. A detailed task analysis as proposed by Gagn6 (1967) or as required for computer simulation often suggests the aptitudes called into play. These three approaches focus with different emphases upon the three components of ATI, i.e., aptitudes, treatments, and tasks. One outcome of this multi-pronged interest is the spotty distribution of findings available to plug into our hypothetical matrix. Studies which might be useful to practitioners are scattered throughout several literatures. The
AN OVERVIEW OF PROBLEMS IN ATI RESEARCH
ATI Components: Aptitudes
:
231
diversity of purpose, the lack of common terminology, and the dissimilarity of level of concern, contribute to the inaccessibility of the information to the would-be implementer. What are the problems posed by the ATI components? 1. What are aptitudes? No clear signals direct the researcher toward a narrow conception of aptitudes. In looking for the differences among individuals which will determine their success on learning tasks, one can consider at least the following: age, social status, mental abilities, sex, personality traits, entering competencies, cognitive styles, values, preferences, motivations, and the individual's perception of the situation. Within most of these categories several variables have been identified as influential in certain situations under particular conditions. None of these aptitudes have been shown to be consistently involved in all treatments and tasks with which they were tested. Exploring "general mental ability" as an exemplum of one of these aptitude categories, we are confronted by numerous internal problems. Measures of general ability of achievement produce a good number of interactions with various treatments and appear to be closely related to "ability to learn" (Cronbach & Snow, in press). Knowing this, however, has not facilitated the development of treatments to serve individuals who have more difficulty in learning. Innovative treatments too often raise the level of performance of highability children while low-ability children respond as before to traditional treatments. This phenomenon increases the achievement gap between children rather than reducing it. Although general ability, as measured on Stanford-Binet or Wechsler tests, has been the predominant conception of cognitive ability employed in ATI research, there is no consensus on what general ability is, or that general ability is sufficient to explain cognitive ability. Guilford (1967) proposes a 120-cell cube consisting of the mental operations, contents, and products which combine to determine an individual's ability on particular tasks. These cells may be tapped individually with measures Guilford has developed, and scores then can be used to predict performance on related tasks. Others skeptical of the validity of these distinctions prefer to regard the structure of intelligence as a hierarchy in
AV C O M M U N I C A T I O N REVIEW 1975 SUMMER
:
232
which a broad general ability underlies more restricted abilities such as spatial reasoning, verbal skills, numerical reasoning, and mechanical understanding. These abilities can be further partitioned into specific skills such as semantic competence, facility with syntax, knowledge of geometry, etc. Each of these skills can again be divided into smaller elements such as knowledge of particular geometric axioms (see Cronbach, 1970). The most productive level of the hierarchical model at which "aptitude" measurements should be made has not been empirically determined. Problem solving approaches suggest still another conception of cognitive abilities worthy of investigation. Based on current knowledge of mental processing of information, measures of individual differences in perception, encoding, storage, and decoding qualify as aptitude indicators. We have, therefore, located at least three distinct conceptualizations of mental ability differences among individuals which could be used in aptitude testing. Other aptitude categories such as cognitive styles and personality traits are bursting with a comparable proliferation of variables. Within "cognitive style" we generally find dichotomies such as field dependent vs. field independent; leveler vs. sharpener; impulsive vs. reflective; and abstract vs. concrete (see Kagan & Kogan, 1970, for a comprehensive review). The aptitudes considered within "personality traits" range from anxiety to authoritarianism, from introversion to need for achievement. While personality variables often produce the strongest interactions with treatments (Cronbach & Snow, in press), they are also often elusive, open to multiple interpretations, and possibly always changing. 2. How do we maintain records of aptitudes? The diversity of aptitudes, all of which may be relevant to some forms of instruction, provide the grist for many educational research mills. However, keeping in mind the ostensible goal, i.e., improvement of the instructional system, we are overwhelmed by numbers: a. On how many aptitudes must each child be measured in order for an instructor to match a student to the appropriate treatment? b. Are we satisfied with one test per aptitude or do we require multi-method validation of scores?
A N OVERVIEW OF PROBLEMS IN ATI RESEARCH
ATI Components: Treatments
:
233
c. How frequently should each child be retested on each aptitude? Do scores change daily? monthly? yearly? ever? The problem of continued aptitude retesting in order to ascertain current standing of students has been tackled only with techniques provided in programed instruction and CAI in which the entering competencies and outcomes of the preceding instructional task are used as aptitude measures to determine the next instructional sequence. 3. H o w do we manage A T I ? Though some remain skeptical about the existence of interactions, we are concerned more with the abundance of interactions. Not cnly do a multitude of aptitudes exist, but it is more than likely that they interact with each other. Sudies in which anxiety, mental ability, and treatment interact are not unheard of. But the investigation of the realistic possibility of higher-order interactions, such as between socioeconomic status, anxiety, mental ability (according to some measure), field dependency, age, past experience, and treatment is more than theory, methodology, or sample size could support. 1. W h a t are treatments? Treatment variables are as plentiful as aptitude factors. The simplest way to organize the treatment universe is by differentiating the vehicles through which instruction is carried. Most frequently encountered are teachers, peers, radio, books, slides, films, broadcast or closed-circuit television, records, tape recordings, programs, and computers. Instructional elements of treatment vehicles are subject to numerous internal differentiations. Within a computer-assisted instructional treatment, for example, decisions must be made as to step size, type and amount of student response demanded, amount and direction of branching, level and order of concept presentation, ratio and order of rules and examples, number and placement of illustrations, and teletype versus CRT presentation. The continual creation of new technologies suggests an infinite set of treatment possibilities. Attempts to reduce this overwhelming variety focus on the unique attributes of vehicles or media which can be related to mediating psychological processes essential to task learning. Broadly defined, an attribute is any structural component of a medium that influences the kind of material one can present, the arrangement of material, and the relation to
AV C O M M U N I C A T I O N
REVIEW 1975 S U M M E R
:
234
other material or the manner in which material is presented (Snow & Salomon, 1968). Salomon is interested in the modes of presentation which supplant or arouse relevant mental operations. In particular, he stresses supplantation, by which he means the execution of the mental process for the individual. This can aid the learner by compensating for an inability, or by creating an image or transformation which the learner can store internally for future use (Salomon, 1970). Research on treatments is moving away from attempts at broad comparisons (e.g., traditional vs CAI) toward the development of taxonomies of attributes which may be implemented by program designers (Bretz, 1971; Tosti & Ball, 1969; Gropper & Glascow, 1971; Fleming, 1967). At present there is little agreement on a categorization scheme for taxonomies and each taxonomy presents a different picture. The nature of the treatment also varies with the context in which a medium is to operate. Differer.ces in atmosphere, structure, discipline, and student freedom and participation may be considered elements of the treatment. 2. H o w s h o u l d w e use treatments? Cronbach and Snow (in press) suggest three ways in which treatments may be designed to assist the learner. In the first case, treatments capitalize on the learner's assets, either on abilities previously built up or on expressed preferences. Although some studies indicate that performance is not significantly better on preferred treatments, students report feeling better about the experience. If ~/ "love of learning" is something we want to inculcate in students, it appears reasonable to respect student choices. This is particularly relevant for the younger students who are developing attitudes toward education and for older students in adult education courses who feel no external pressures to continue learning. However, treatments dependent upon previously attained abilities may result in routing students repeatedly into one type of presentation. Although the learner develops competence in subject matters treated with the familiar techniques, he or she may be inept when confronted with unfamiliar treatments. Cronbach and Snow cite one study in particular (Browell & Moser, 1949) in which children who had had "rote" instruction in subtraction for two years could not incorporate the meanings of the mathematical operations
AN OVERVIEW OF PROBLEMS IN ATI RESEARCH
:
235
when they were offered. Danger arises here if knowledge of a student's aptitudes determines a like treatment across tasks to the neglect of less developed aptitudes or abilities which also have potential. This approach to treatment design is less threatening for mature students who have developed strengths over a long period of time and will optimize learning by relying on these aptitudes. The second strategy of treatment design provides the compensation to which Salomon refers. Thu~, for the child with poor visual abilities, the treatment provides images. The danger inherent in this approach, if it is repeated consistently, is the development of reliance upon the compensatory techniques rather than the development of the ability. The introduction of a pocket-sized calculatc, r has increased the saliency of this issue. Current debate revolves around the use of calculators in the classroom or for homework. Do we dare allow children to use calculators before they have developed a proficiency in basic mathematical skills? Aptitudetreatment interaction research has rarely entertained child development issues. However, treatments designed to optimize on a known aptitude may be desirable only after a child has had the opportunity to develop a wide variety of responses. Perhaps compensatory techniques should not be used for students until is has been established that they will not accomplish the task without such assistance. The third approach suggested by Cronbach and Snow is remediation, in which intervention assesses the underdeveloped aptitudes and attempts to develop the specific skills which are lacking. Connections are made so that the individual can perform the operation independently at a later time. 3. Getting treatments into the real world. Generally, as we mentioned above, people designing treatments are familiar with one or two media, using them to furnish and to test their ideas. Rarely does a researcher have the sufficiently broad competence necessary for multimedia comparisons so that optimal treatments can be designed. Treatment implementation poses additional obstacles. Assuming the possibility of designing one treatment that heightens the performance of individuals high on certain aptitudes and another treatment that enhances the performance of students low on those aptitudes, the question arises
AV C O M M U N I C A T I O N REVIEW 1975 SUMMER
:
2,36
as to how to make the appropriate treatment available to students. It is unrealistic to expect that each school could afford the variety implied, yet this is the logical extension of any argument for aptitude-treatment interaction research. Unless deliberately diverted, the distribution of treatment innovation may well follow the present pattern which disproportionately enriches programs of wealthy school districts. The costs of individualized treatments are bound to be high for commercial producers. Although educational researchers do not often consider market costs, the eventual application of the treatments they develop depends on financial feasibility. The market compels producers to prepare materials appropriate for the largest number of consumers rather than for a select group. Economic pressure suggests flexible treatments which can be adapted by learners with different needs and abilities, rather than specialized treatments which aim at one specific segment of the population. While the advantages reaped from CAI are not yet equal to the enormous investment costs, the relatively low cost of modifying software should eventually make it attractive to even the less affluent school districts. ATI Components: Tasks
1. W h a t are tasks? Whereas researchers and producers begin
with an aptitude or medium in mind, the practitioner generally focuses on a task. The student's performance on learning tasks is used as a criterion of educational success and teacher accountability. The range of tasks is extremely broad. In public schools, instructional material is divided into subjects and courses. Each course is subdivided into the particular skills and knowledge necessary to attain competence in and comprehension of the subject. The learning of skills and knowledge is then broken down into the steps required for mastery. Some tasks such as pattern recognition, memorization, and organization generalize across content areas, whereas others such as matching the names of presidents to their home states or bisecting an angle are content specific. Research in areas that have yielded inconsistent results (as in the programed instruction studies mentioned above) frequently ignore differences in subject "content." It is not clear that learning statistics through programed instruction and learning Spanish through programed instruction are comparable
AN OVERVIEW OF PROBLEMS IN ATI RESEARCH
:
237
tasks. Even two independently produced statistics PI courses may be presenting different tasks to the students. One program might stress derivation of equations whereas the second might emphasize application of the equations. Detailed descriptions of such tasks are generally omitted from research reports concerned with ATI. The resulting ambiguity detracts from the use of summary reports for future treatment designers or implementers.1 A further obstacle in task definition emerges as we consider the various goals educators and parents set for education. As the values and objectives vary, the tasks and skills students a r e expected to master will change. Whether an instructional system focuses on emotional development, preparation for the job market, or problem-solving ability will determine the level and range of tasks a student must accomplish. The difficulty of narrowing the field of tasks so that our hypothetical matrix can be sketched in is again increased by the creation of new technologies. Each new medium presents an array of new learning tasks as well as new treatment possibilities. In addition, use of the new technology may bring to light aptitudes which were previously unmeasured. The educational system, however, has been slow to absorb innovations not deliberately placed within its realm. A case in point is television. Thus far, although students by age 16 have spent as much time in front of television as they have in school (Roberts, 1973; Liebert, Neale, & Davidson, 1973), they are given no instruction in media literacy, criticism, or television production techniques. Television is considered outside the jurisdiction of schools; new task areas presented by the new medium are ignored. When television does find its way into the classroom, it is only as the vehicle for conveying conventional instructional information. The problems posed by particular task areas can only be considered once the infinite array of possible tasks has been limited in some manner.
lLumsdaine (1972) has drawn a distinction between the two "faces" of One," what is presented-i.e., the objectives or intended outcomes of the program-is compatible with the meaning in this section. Lumsdaine's "Content Two," or how it is presented, is closer to the considerations under "'treatments" in this paper. content: "Content
AN COMMUNICATION REVIEW 1975 SUMMER
IMPLICATIONS FOR FUTURE RESEARCH
:
238
Several "real-world" factors must be weighed in determining the future directions of aptitude-treatment interaction research in instruction. First, resources devoted to research are more scarce than they were in the 1960s. As educational researchers compete for funding, they should be aware that the opportunity costs of producing a particular educational program are measured across all sectors (e.g., health care, defense, energy) of society rather than solely within the educational sector. If individual researchers continue to work independently, producing results which are never integrated or applied, funding will disappear (as the current fortunes of the National Institute of Education illustrate). To prevent a congressionallyimposed starvation diet, instructional researchers must determine, develop, and deliver a service or good desired by the market. The appropriate marketplace for educational research is the instructional situation. It is here that the educational researcher hopes eventually to have an impact, and it is here, finally, that his or her work is evaluated. In the past, however, the pathway from the researcher's laboratory to the classroom has often been circuitous if not totally blocked. To ensure a more direct contribution, ATI researchers might begin with the instructional task. Several authors have suggested task analysis as the origin point for ATI research (Rhetts, 1974; Cronbach & Snow, in press; Gagn6, 1967). The products of educational research must be accessible to the practitioner as well as applicable in the "real-world." Although this has proved difficult in the past, information technologies (e.g., large data banks and computer terminals in the schools) will change the organization and dissemination possibilities for research findings. When research results are available to practitioners at their place of work (i.e., in the schools), they will more likely be used. These information technologies will also make communication from the practitioner to the researcher possible. Rarely in the past have practitioners had access to researchers. They have had no opportunity to raise questions they need answered; nor have they been able to share their experiences with researchers. It may be that the researcher has much to learn from the daily insights of the concerned instructor.
AN OVERVIEWOF PROBLEMSIN ATI RESEARCH :
CONCLUSION
REFERENCES
239
Though not scientifically valid, the input from practitioners may point the direction for future ATI research. Mindful of the problems encountered in ATI research and the political, economic, and technological context within which future work will be undertaken, we suggest a more deliberate path than has been pursued in the past. A concerted effort by ATI researchers which focuses on a few tasks rather than on the whole matrix is more likely to demonstrate whether this research has a significant contribution to make to instructional practices. Drawing researchers together will encourage content and concept validation and conformity to methodological standards. The inclusion of individuals with diverse backgrounds and expertise will facilitate comparisons within and across treatments while maintaining aptitudes and tasks across researchers. The participation of practitioners can be increased as access to computer terminals in libraries and schools becomes more widespread. If the trend toward the individualization of instruction for the student is to continue, it may be necessary for researchers to surrender a fraction of their own individuality to the team effort. Allen, W. H. Instructional media research: Past, present and future. A V Communication Review, 1971, 19, 9-18. Ball, B., & Bogatz, G. A. A summary of the major findings in "The first year of Sesame Street: An evaluation." Princeton, N.J.: Educational Testing Service, 1970. Bretz, R. A Taxonomy of communication media. Englewood Cliffs, N.J.: Educational Technology Publications, 1971. Campbell, D. T., & Fiske, D. W. Convergent and discriminant validation by the mulfitrait-multimethod matrix. Psychological Bulletin, 1959, 56, 81-105. Campeau, P. L. Selective review of the results of research on the use of audiovisual media to teach adults. AV Communication Review, 1974, 22, 5-40. Coleman, J., et al. Equality of educational opportunity. Washington, D.C.: U.S. Department of Health, Education and Welfare, Office of Education, 1966. Cronbach, L. J. The two disciplines of scientific psychology. American Psychologist, 1957, 12, 671-684. Cronbach, L. J. Essentials of psychological testing. (3rd ed.) New York: Harper and Row, 1970. Cronbach, L. J., & Snow, R. E. Individual differences in learning ability as a function of instructional variables. Final Report.
AN COMMUNICATIONREVIEW1975 SUMMER :
240
USOE Contract No. OEC4-6-061269-1217. Stanford University, 1969. Cronbach, L. J., & Snow, R. E. Aptitudes and instructional methods. New York: Irvington, in press. Fleming, M. Classification and analysis of instructional illustrations. A V Communication Review, 1967, 15, 246-258. Gagn~, R. (Ed.) Learning and individual differences. Columbus, Ohio: Charles E. Merrill, 1967. Gropper, G. U & Glasgow, Z. Criteria for the selection and use of visuals in instruction: A handbook. Englewood Cliffs, N.J.: Educational Technology Publications, 1971. Guilford, J. P. The nature of human intelligence. New York: McGraw-Hill, 1967. Jamison, D., Suppes, P., & Wells, S. The effectiveness of alternative instructional media: A survey. Stanford, Calif.: Stanford University, The Graduate School of Business, 1973. Kagan, J., & Kogan, N. Individuality and cognitive performance. In P. Mussen (Ed.), Carmichael's manual of child psychology. (3rd ed.) New York: John Wiley, 1970. Pp. 1273-1365. Koran, M. L., Snow, R., & McDonald, F. J. Teacher aptitude and observational learning of a teaching skill. Journal of Educational Psychology, 1971, 62, 219..228. Levie, H. W., & Dickie, K. E. The analysis and application of media. In R. M. W. Travers (Ed.), Second handbook of research on teaching. Chicago: Rand McNally, 1973. Pp. 858-882. Liebert, R. M., Neale, J. M., & Davidson, E. S. The early window. New York: Pergamon, 1973. Lumsdaine, A. A. "Content" and the outcomes of educational programs. In W. Schramm (Ed.), Quality of instructional television. Honolulu: University Press of Hawaii, 1972. Merrill, M. D. Learner control: Beyond aptitude-treatment interactions. A V Communication Review, 1975, 23, (2), 217-226. Mielke, K. W. Asking the right ETV research questions. Educational Broadcasting Review, 1968, 2(6), 54-61. Mielke, K. W. Media-message interactions in TV. In G. Salomon & R. E. Snow (Eds.), Viewpoints (Bulletin of the School of Education, Indiana University), 1970, 46, 15-31. Rhetts, J. E. Task, learner and treatment variables in instructional design. Journal of Educational Psychology, 1974, 66(3), 339347. Roberts, D. F. Communication and children: A developmental approach. In I. D. Pool & W. Schramm (Eds.), Handbook of communications. Chicago: Rand-McNally, 1973, Pp. 174-215. Salomon, G. What does it do to Johnny? A cognitive-functionalistic view of research on media. In G. Salomon & R. E. Snow (Eds.), Viewpoints (Bulletin of the School of Education, Indiana University), 1970, 46, 33-62. Salomon, G. Can we affect cognitive skills through visual media?
AN OVERVIEWOF PROBLEMSIN ATI RESEARCH :
241
An hypothesis and initial findings. AV Communication Review, 1972, 20, 401-422. Salomon, G. What is learned and how it is taught: The interaction between media, message, task, and learner. In D. Olson (Ed.), NSSE yearbook: Media and symbols, the forms of expression, communication and education. Chicago: The University Press, 1974. Schramm, W. What the research says. In W. Schramm (Ed.), Quality in instructional television. Honolulu: The University Press of Hawaii, 1972. Snow, R. E. Research on media and aptitudes. In G. Salomon & R. E. Snow (Eds.), Viewpoints (Bulletin of the School of Education, Indiana University), 1970, 46, 63-89. Snow, R. E., & Salomon, G. Aptitudes and instructional media. A V Communication Review, 1968, 16, 341-357. Tosti, D. T. & Ball, J. R. A behavioral approach to instructional design and media selection. AV Communication Review, 1969, 17, 5-25.