1) The document discusses the importance of controlling for various variables when evaluating the effectiveness of instructional programs. These variables include student aptitude, support factors, and how the instructional program was implemented (process variables).
2) It explains different methods for controlling these variables, such as randomly assigning students, matching schools based on student aptitude, and using statistical methods to account for variable effects.
3) Controlling for these external influences is necessary to isolate the "net effect" of the instruction and obtain valid evidence about how well the program achieved its objectives or how it compares to other programs. Failing to control for variables could lead to incorrect conclusions about a program's effectiveness.
1) The document discusses the importance of controlling for various variables when evaluating the effectiveness of instructional programs. These variables include student aptitude, support factors, and how the instructional program was implemented (process variables).
2) It explains different methods for controlling these variables, such as randomly assigning students, matching schools based on student aptitude, and using statistical methods to account for variable effects.
3) Controlling for these external influences is necessary to isolate the "net effect" of the instruction and obtain valid evidence about how well the program achieved its objectives or how it compares to other programs. Failing to control for variables could lead to incorrect conclusions about a program's effectiveness.
1) The document discusses the importance of controlling for various variables when evaluating the effectiveness of instructional programs. These variables include student aptitude, support factors, and how the instructional program was implemented (process variables).
2) It explains different methods for controlling these variables, such as randomly assigning students, matching schools based on student aptitude, and using statistical methods to account for variable effects.
3) Controlling for these external influences is necessary to isolate the "net effect" of the instruction and obtain valid evidence about how well the program achieved its objectives or how it compares to other programs. Failing to control for variables could lead to incorrect conclusions about a program's effectiveness.
We have pointed out that measures of the outcomes of an instructional pro-
gram—that is, measures of learned intellectual skills, cognitive strategies, information, attitudes, and motor skills—are influenced by a number of variables in the educational situation besides the program itself. Process variables in the operation of the instructional program may directly affect learning and, thus, also affect its outcomes. Support variables in the school or in the home determine the opportunities for learning and, thus, influence the outcomes of learning that are observed. And most prominently of all, the learning aptitude of students strongly influences the outcomes measured in an evaluation study. If the effectiveness of the designed instruction is to be evaluated, certain controls must be instituted over process, support, and aptitude variables to ensure that the "net effect" of the instruction is revealed. Procedures for accomplishing this control are described in this section. Again, it may be necessary to point out that only the basic logic of these procedures can be accounted for here. However, such logic is of critical importance in the design of evaluation studies. Controlling for Aptitude Effects The assessment of outcomes of instruction in terms of question 1 (To what extent have objectives been met?) needs to take account of the effects of aptitude variables. In the context of this question, it is mainly desirable to state what is the level of intelligence of the students being instructed. This may be done most simply by giving the average score and some measure of dispersion of the distribution of scores (such as the standard deviation) on a standard test of intelligence. However, correlated measures such as SES are frequently used for this purpose. Supposing that 117 out of 130 objectives of a designed course are found to have been met, it is of some importance to know whether the average IQ of the students is 115 (as might be true in a suburban school) or 102 (as might occur in some sections of a city or in a rural area). It is possible that, in the former setting, the number of objectives achieved might be 117 out of 130, whereas in the latter, this might drop to 98 out of 130. The aims of evaluation may best be accomplished bv trving out the instructional entity in several different schools, each having a somewhat different range of student learning aptitude. When the purposes of question 2 (To what degree is it better?) are being served in evaluation, one must go beyond simply reporting the nature and amount of the aptitude variable. In this case, the concern is to show whether any difference exists between the new instructional program and some other—in other words, to make a comparison. Simply stated, making a comparison requires the demonstration that the two groups of students were equivalent to begin with. Equivalence of students in aptitude is most likely to occur when successive classes of students in the same school, coming from the same neighborhood, are employed as comparison groups. This is the case when a newly designed course is introduced in a classroom or school and is to be compared with a different course given the previous year. Other methods of establishing equivalence of initial aptitudes are often employed. Sometimes, it is possible to assign students randomly to different classrooms within a single school, half of which receive the newly designed instruction and half of which do not. When such a design is used, definite administrative arrangements must be made to ensure randomness—it cannot be assumed. Another procedure is to select a set of schools that are "matched," insofar as possible, in the aptitudes of their students and to try out the new instruction in half of these, making a comparison with the outcomes obtained in those schools not receiving the new instruction. All of these methods contain certain complexities of design that necessitate careful management if valid comparisons are to be made. There are also statistical methods of control for aptitude variables—methods that "partial out" the effects of aptitude variables and, thus, reveal the net effect of the instruction. In general, these methods follow this logic: If the measured outcome is produced by A and I, where A is aptitude and I is instruction, what would be the effect of I alone if A were assumed to have a constant value rather than a variable one? Such methods are of considerable value in revealing instructional effectiveness, bearing in mind particularly the prominent influence the A variable is likely to have. Whatever particular procedure is employed, it should be clear that any valid comparison of the effectiveness of instruction in two or more groups of students requires that equivalence of initial aptitudes be established. Measures of intelligence, or other correlated measures, may be employed in the comparison. Students may be randomly assigned to the different groups or their aptitudes may be compared when assignment has been made on other grounds (such as school location). Statistical means may be employed to make possible the assumption of equivalence. Any or all of these means are aimed at making a convincing case for equivalence of learning aptitudes among groups of students whose capabilities following instruction are being compared. No study evaluating learning outcomes can provide valid evidence of instructional effectiveness without having a way of controlling this important variable. Controlling for the Effects of Support Variables For many purposes of evaluation, support variables may be treated as input variables and, thus, controlled in ways similar to those used for learning aptitude. Thus, when interest is centered upon the attainment of objectives (question 1), the measures made of support variables can be reported along with outcome measures so that they can be considered in interpreting the outcomes. Here again, a useful procedure is to try out the instruction in a variety of schools displaying different characteristics (or different amounts) of support. Similarly, the comparisons implied by question 2 and part of question 3 require the demonstration of equivalence among the classes or schools whose learning outcomes are being compared. Suppose that outcome measures are obtained from two different aptitude-equivalent groups of students in a school, one of which has been trying out a newly designed course in English composition, while the other continues with a different course. Assume that, despite differences in the instruction, the objectives of the two courses are largely the same and that assessment of outcomes is based on these common objectives. . Class M is found to show significant^ better performance, on the average, than does class N. Before the evidence that the new instruction is "better" can be truly convincing, it must be shown that no differences exist in support variables. Since the school is the same, many variables of this sort can be shown to be equivalent, such as the library, the kinds of materials available, and others of this nature. Where might differences in support variables be found? One possibility is the climate of the two classrooms— one may be more encouraging to achievement than the other. Two different teachers are involved—one may be disliked, the other liked. Student attitudes may be different—more students in one class may seek new opportunities for learning than do students in the other. Variables of this sort that affect opportunities for learning may accordingly affect outcomes. Therefore, it is quite essential that equivalence of groups with respect to these variables be demonstrated or taken into account by statistical means. Controlling for the Effects of Process Variables The assessment and control of process variables is of particular concern in seeking evidence bearing on the attainment of stated objectives (question 1). Quite evidently, an instructional entity may work either better or worse depending upon how the operations it specifies are carried out. Suppose, for example, that a new course in elementary science presumes that teachers will treat the directing of students' activities as something left almost entirely to the students themselves (guided by an exercise booklet). Teachers find that under these circumstances, the students tend to raise questions to which they (the teachers) don't always know the answers. One teacher may deal with this circumstance by encouraging students to see if they can invent a way of finding the answer. Another teacher may require that students do only what their exercise book describes. Thus, the same instructional program may lead to quite different operations. The process variable differs markedly in these two instances, and equally marked effects may show up in measures of outcome. If the evaluation is of the formative type, the designer may interpret such evidence as showing the need for additional teacher instructions or training. If summative evaluation is being conducted, results from the two groups of students must be treated separately to disclose the effects of the process variable. In comparison studies (question 2), process variables are equally important. As in the case of aptitude or support variables, they must be controlled in one wav or another in order for valid evidence of the effectiveness of instruction to be obtained. Equivalence of groups in terms of process variables must be shown, either by exercising direct control over them by a randomizing approach, or by statistical means. It may be noted that process variables are more amenable to direct control than are either support or aptitude variables. If a school or class is conducted in a noisy environment (a support variable), the means of changing the noise level may not be readily at hand. If, however, a formative evaluation study shows that some teachers have failed to use the operations specified by the new instructional program (a process variable), instruction of these teachers can be undertaken so that the next trial starts off with a desirable set of process variables. Unanticipated outcomes (question 3) are equally likely to be influenced by process variables and accordingly require similar control procedures. A set of positive attitudes on the part of students of a newly designed program could result from the human modeling of a particular teacher and, thus, contrast with less favorable attitudes in another group of students who have otherwise had the same instruction. It is necessary in this case, also, to demonstrate equivalence of process variables before drawing conclusions about effects of the instructional entity. Controlling Variables by Randomization It is generally agreed that the best possible way to control variables in an evaluation study is to ensure that their effects occur in a random fashion. This is the case when students can be assigned to control and experimental groups in a truly random manner or when an entire set of classes or schools can be divided into such groups randomly. In the simplest case, if the outcomes of group A (the new instructional entity) are compared with those of group B (the previously employed instruction), and students drawn from a given population have been assigned to these groups in equal numbers at random, the comparison of the outcomes may be assumed to be equally influenced by aptitude variables. Similar reasoning applies to the effects of randomizing die assignment of classrooms, teachers, and schools to experimental and control groups in order to equalize process and support variables. Randomization has the effect of controlling not only the specific variables that have been identified, but also other variables that may not have been singled out for measurement because their potential influence is unknown. Although ideal for purposes of control, in practice, randomizing procedures are usually difficult to arrange. Schools do not customarily draw their students randomly from a community or assign them randomly to classes or teachers. Accordingly, the identification and measurement of aptitude, support, and process variables must usually be undertaken as described in the preceding sections. When random assignment of students, teachers, or classes is possible, evaluation studies achieve a degree of elegance they do not otherwise possess.
Empath: The Survival Guide For Highly Sensitive People: Protect Yourself From Narcissists & Toxic Relationships. Discover How to Stop Absorbing Other People's Pain
Raising Mentally Strong Kids: How to Combine the Power of Neuroscience with Love and Logic to Grow Confident, Kind, Responsible, and Resilient Children and Young Adults
Dark Psychology: Learn To Influence Anyone Using Mind Control, Manipulation And Deception With Secret Techniques Of Dark Persuasion, Undetected Mind Control, Mind Games, Hypnotism And Brainwashing
Dark Psychology & Manipulation: Discover How To Analyze People and Master Human Behaviour Using Emotional Influence Techniques, Body Language Secrets, Covert NLP, Speed Reading, and Hypnosis.