Developing an Understanding of the
Role of Feedback in Educationby Shani M DOIG
Griffith University
Introduction
The important role of feedback in improving student performance has long been recognised by educational researchers (Mory, 1992). There is still considerable variation, however, regarding the provision of purposeful feedback to students. While the term feedback still remains problematic within the domain of teaching, learning and assessment, scientific uses of the term are considerably clearer. An exploration of the uses of the term feedback in science, particularly in process-control systems, show it as being defined and contained within particular system parameters and ultimately linked to system objectives and outcomes. Obviously, the certainty of science cannot be replicated in any activity involving human agents and hence the transferability of scientific models of feedback to educational contexts is limited. This is not to say, however, that scientific models should be dismissed out of turn as having no benefit. On the contrary, an exploration of the uses of feedback in science can provide valuable insights that can further understandings of the way in which feedback is considered and used within educational settings. This paper, in the first instance, aims to explore the use of the term feedback in science, through a discussion of process control systems. It then provides illustrative instances of how the term has been adopted by educational theorists working in the fields of cognition, mastery learning and assessment. The aim of the paper is not to suggest a 'rightness' or 'wrongness' in the way the term feedback has come to be discussed and defined in education, but rather, to seek to explore possible areas in the field that are worthy of further exploration and research. Specifically, it is argued, that the role of pedagogy within the feedback process needs to be the subject of further sustained, theoretical exploration.
Feedback and Process Control Systems
The term feedback has a recent history, coming into noted usage in the 1930s to describe a function of process-control systems. Such systems may be defined as systems in which "deliberate guidance or manipulation is used to achieve a prescribed value of a variable" (Scientific Apparatus Makers Association, 1970, p. 11). In such systems feedback is used primarily as an error detection mechanism and functions to provide guidance to the system, such that is can operate effectively. In order for process-control systems to function it is necessary that what is happening in the system at a given time is 'fed back' into the array and compared to some standard of what should be happening. The thermostat, for example, performs this feedback function with respect to heating systems. In such systems, the temperature of a room at a given time is noted and this information is fed back into the heating system, such that the unit is turned on when the room temperature falls below a given value and turned off when the room is sufficiently warm. Thus the term feedback is defined within the field of process-control systems as being a:
Control action in which a measured variable is compared to its desired value to produce an actuating error signal which is acted upon in such a way as to reduce the magnitude of the error. (Scientific Apparatus Makers Association, 1970, p. 8)The process of feedback may be seen then as having three key elements. In the first instance the system must be working towards a known objective (the heating of a room to a given temperature). Secondly, the system, in operation, must be monitored in such a way that any diversion from this objective on the part of the system is able to be detected (when the room is too hot or too cold) and thirdly, any gap between what should be happening and what is happening is able to be rectified by adaptations being made to the operation of the system (the heater is turned off or on). Feedback then does not operate in an isolated capacity, but rather exists within a tightly defined and controlled system in which objectives and outcomes are known from the outset. Within the contexts of such a system, feedback operates to assist with the regulation of the system by detecting any deviance on the part of the system that if uncorrected would result in the system failing to meet its required objectives or produced its desired outcome. In order then for feedback to work effectively in process control systems it is necessary that both the objectives and outcomes required by the system be clearly defined and that the feedback provided to the system correspond to these preset objectives and outcomes. For example it would be pointless for a thermostat to feedback information into a heating unit which had not been programmed to heat a room to a certain temperature. Likewise, the system would be flawed if, having been programmed to heat a room to a certain temperature, the unit was incapable of performing the required action when the feedback suggested that the temperature was not being achieved.
Feedback, as it relates to process control systems has three elements, which are:
- That it works within a controlled and defined system;
- That the objectives and desired outcomes of the system are known from the outset;
- That the feedback provided is utilised within the contexts of the system such that necessary changes are made.
The maintenance and regulation of these three features are not always easy even in mechanistic systems, as recent difficulties within Australian service industries glaringly demonstrate. Assuredly one of the problems lies in the fact that science cannot readily describe or mandate the ways in which human agents will impact upon the operation of mechanical systems. Defining and describing feedback becomes even more problematic when discussion turns to its use in a field such as education, being as it is entirely constituted of human agents.
From Feedback in Science to Feedback in Education: Problems and Possibilities
As has already been alluded to, scientific systems have regimented and defined parameters whereas social systems, on the other hand, often defy the application and rigid definitions and the setting of exact limits. According to Hughes (1958):
The essential difference between the natural and the human world, [Weber] contended, lay in the fact that in the latter realm it was impossible to arrive at law - or causal explanations - that would in any sense give a satisfactory or exhaustive explanation of even the simplest of human actions. (in Worsley, 1985, p. 25)Thus it must be considered that there can be no simplistic transference of scientific definitions about feedback to the social context of the educational sphere. Mechanistic definitions of the term may provide some guidance but they cannot be exactly replicated in educational contexts - populated as they are by human agents involved in institutional settings and complex power relations.
This is a powerful reason why scientific uses of the term feedback have often been omitted in discussions relating to its provision in educational contexts. Even at its broadest level, for example, the notion of feedback as being defined according to the system of which it is part becomes problematic when transferred to an educational paradigm. The question of what exactly constitutes an educational system yields no readily manageable answer. Rather, the definition of an educational system is by necessity multi-layered and incorporates elements of curriculum, pedagogy, policy and theory, not to mention the amalgous entity that is 'the student'. Each of these components can and should be broken down into numerous further subsets. Concomitantly, the explication of system objectives and outcomes is by no means an easy task as these to can be defined from a number of different perspectives dependent upon where one is positioned within the complex matrix of the educational system and according to the relative power position that one holds within that system.
The difficulties in transcending the fields of science and education are perhaps best exemplified by Elshout-Mohr (1994) who begins her discussion of feedback in education with an attempted definition of feedback in the field of cybernetics ( a field closely related to process-control systems). She states that in this field feedback:
...refers to the regulatory mechanisms which, for example, make it possible to decorate a whipped-cream cake in a moving train using a spray tube that does not put out an even quantity of cream at every moment. The decoration is not supposed to win a beauty prize, but it will probably exceed what one might expect if feedback did not exist. For feedback, it is necessary for the acting person to have an eye on the result of what he or she does (the cake and the spray) and that he or she is "so made" that the regulatory mechanism works. (p. 58/59)This explanation of feedback in a field of science is limited, in the first instance, by its brevity. In attempting to encapsulate how feedback operates within the field of cybernetics, the discussion by Elshout-Mohr (1994) either skims over essential elements of the term or demonstrates a significant lack of concern with issues which are not only of vital concern with respect to providing a scientific definition of the term but are also necessary considerations when attempting a related discussion of the uses of feedback in education. While an attempt is made within the extract to define feedback in relation to an understood objective - the decoration of a cake on a moving train - there is no sense of how these objectives are defined within a clearly understood and defined system. Similarly, how the nature of the feedback will be determined by the desired characteristics of the finished product are implicit, rather than explicit, within the discussion provided. Consequently there is no discussion of how the progress of the cake decoration will be monitored and fed back into the system causing action to either be altered or maintained such that the final outcome is achieved. There is also no recognition that there is an inherent connection between the standard to be reached by the finished product and the feedback provided during its production. There is no sense of whether or not the feedback would change if the cake was expected to win the beauty contest. Finally, the extract neither defines nor explains the term "so made", a concept which would appear to be a vital linchpin in developing an understanding of the way in which feedback comes to be utilised in such a situation as the one presented. The difficulties inherent in a simple transference from the explanation of feedback provided in science to an educational context, may explain why few theorists in education have chosen to directly discuss the scientific origins of the word. Instead, there has been an attempt across a number of fields of educational theory to explore only the practice of feedback as it relates to teaching, learning and assessment.
Understandings of the Role of Feedback in Education
What follows are illustrative instances of the ways in which the nature and function of feedback in education has been considered. The examples provided, coming from the fields of cognition, mastery learning and assessment, show the complexity involved in defining and describing the role of feedback in education.
Viewing the issue of feedback from a cognitive perspective, Butler and Winne (1995) examine how feedback comes to be understood by students according to their individual learning objectives. In a sense the article deals with one of the issues raised previously with respect to the Elshout-Mohr (1994) article, that being how students come to be "so made" as a result of receiving instructive feedback. According to Butler and Winne "feedback describes the nature of outcomes and the qualities of the cognitive processing that led to those states (p. 246). In this definition there is a view the effectiveness of feedback is determined by anticipated outcomes. The 'qualities of cognitive processing' required to achieve these outcomes are further defined as a set of criteria generated by the learner which assist them to monitor their own performance as they work towards the desired outcome.
According to Butler and Winne (1995), feedback can be both internal (performed by the student themselves) or external (generated by a person/persons other than the student ). The article suggests that while teachers may feedback information on student work regularly, this feedback can only be incorporated into student learning when it is sort by them and related specifically to their individual learning goals and objectives. Concomitantly, the argument is made that if a learner does not possess or has not developed goals or objectives with respect to a desired learning outcome than feedback can do little to assist them in the learning process. The pedagogical implications of constructing individually tailored feedback are not discussed in the article. There is also an omission of how wider curriculum parameters and the objectives and desired outcomes created within these, necessarily impact upon the way in which students come to define their own objectives, understand the feedback provided by teachers, and indeed come to define and regulate their own internal feedback.
Not all writers discuss feedback without reference to curriculum issues. Proponents of mastery learning, for example, discuss feedback within a tightly defined and rigidly prescribed system of teaching, learning and assessment. According to William and Black (1998):
Mastery learning originated as a practical implication of the learning theories of John B Carroll. Carroll had proposed that success in learning was a function solely of the ratio of the time actually spent learning to the time needed for learning - in other words, any student could learn anything if they studied it long enough. (p.24)In addition to issues of duration, theories of mastery learning also mandate that in order to learn successfully students need to have a clear understanding of exactly what it is that they are required to learn and to be provided with accurate feedback relating any errors and/or difficulties that arise as they attempt to meet these learning objectives. Thus Slavin (1987) defines mastery learning in the following manner:
The principal defining characteristic of mastery learning methods is the establishment of a criterion level of performance held to represent "mastery" of a given skill or concept, frequent assessment of student progress toward the mastery criterion, and provision of corrective instruction to enable students who do not initially meet the mastery criterion to do so on later parallel assessment. (p. 1)
The need for frequent assessment of student learning is achieved in mastery learning by regular testing. Tests are designed both to identify and diagnose student difficulties and are also used as the basis for providing feedback to students on or about these difficulties. According to Bangert-Drowns, Kulik and Kulik (1991) tests provide information such that "students who don't show mastery have their weaknesses diagnosed, receive corrective instruction and are given new opportunities to show mastery" (p. 90). Within mastery learning, tests are designed both to fulfil a function of error-detection and also to assist students in their understanding of these errors such that they are able to correct them prior to future tests..
Within mastery learning feedback is associated with predetermined objectives and outcomes. Unlike those described by cognitive theorists, however, the objectives and outcomes are not discussed as being the responsibility of individual learners. Rather objectives and outcomes are seen as operating at a curriculum level and are consequently common for all students. The recognition that some students may take longer to reach the mandated objectives and outcomes is a key feature of mastery learning theory.
An omission within the field of mastery learning with respect to feedback would appear to be exactly how testing instruments are utilised within classroom teaching learning and assessment such that they provide effective feedback to students. Rather, there seems to have been an unproblematic assumption that frequent testing can both diagnose student weaknesses and similarly act as a mechanism for providing effective feedback to students about these weaknesses. Some recent research (Bangert-Drowns, Kulik and Kulik, 1991) would tend to counter such claims, suggesting that frequent testing does not necessarily produce meaningful feedback which students can utilise to improve their performance. For the most part, however, the roles of teachers in translating test results and adapting classroom practice to meet student needs seems to have been given little attention by advocates of mastery learning.
Theorists working within the field of assessment have also considered the issue of what constitutes effective feedback (Daws & Singh, 1996; Arthur, 1995; Gipps, 1994; Bernard & Naidu, 1992; McCurdy and Shapiro, 1992; Mory, 1992; Sadler, 1989; McMenniman, 1986). Sadler's (1989) discussion of feedback as it pertains to formative assessment provides an informing and theorised consideration of the general principals of effective educational feedback. Sadler works from Ramprasad's definition of feedback as being "information about the gap between the actual and reference level of a system parameter which is used to alter the gap in some way" (p. 4). It is interesting to note that Ramprasad's definition, originating as it does in the behavioural sciences, discusses feedback with respect to a controlled system. Sadler, while developing points of Ramprasad's argument does not attempt to transfer the concept of system as used by Ramprasad's to the field of education. Rather, Sadler's main concern would seem to lie in another aspect of Ramprasad's definition.
An important feature of Ramprasad's definition is that information about the gap between actual and reference levels is considered as feedback only when it is used to alter the gap. If the information is simply recorded, passed to a third party who lacks either the knowledge or the power to change the outcome, or is too deeply encoded to lead to appropriate action, the control loop cannot be closed and "dangling data" substitutes for appropriate feedback. (p. 121)
Sadler's (1989) primary concern is with how feedback needs to be provided to students such that the situation of "dangling data" can be avoided. At issue also is how information received during the feedback process is able to be utilised by students such that their performance can be improved. In order to do this Sadler advocates that students receive explicit knowledge about and training in the use of the standards on which work is be judged, stating that:
Stated explicitly, therefore, the learner has to (a) possess a concept of the standard (or goal or reference level) being aimed for, (b) compare the actual (or current) level of performance with the standard, and (c) engage in appropriate action which leads to some closure of the gap. (p. 121)Standards, according to Sadler (1989) are then the medium through which students come to identify the difference between the current and the desired level of performance. Similarly, they act as a guide to students with respect to the actions needed to be taken in order to close any gap that may exist. In total, the provision of standards as an instrument of feedback act as a way of "pulling individual students back on target towards the eventual realisation of long term objectives" (McMeniman, 1986, p. 4).
McMeniman (1986) working again in the area of feedback and formative assessment suggests that individual students establish their own learning objectives and that these in turn influence the concept of a 'desired learning outcome'. Sadler (1989), however, would seem to suggest that objectives and outcomes are prescribed at a level beyond the student. He argues, however, that objectives and outcomes when translated into discrete standards can come to be understood and used by students such that they can guide the improvement of performance. Consequently, what Sadler advocates is a transference model where knowledges of what constitutes effective performance are transferred to students such that they gain understanding and ownership of them. What Sadler does not describe, however, is how this transference occurs. While stating that teachers, in the first instance, hold knowledges of what constitutes effective performance and requisite standards, what Sadler does not discuss is the nature of the pedagogy required to feedback this information to students such that they come to possess similar understandings.
Conclusion
In reviewing discussions of feedback drawn from the fields of cognition, mastery learning and assessment, certain similarities can be noted across the three fields and indeed between them and uses of the term made in science. In the first instance, there is common concern both in science and in education that effective feedback must be related to a set of predetermined objectives and outcomes. Educational theorists do not concur, however, as to exactly how these features should be defined in an educational context. Cognitivists, such as Butler and Winne (1995) argue that objectives and outcomes can only be defined by the individual learner. Advocates of mastery learning (Slavin, 1987), on the other hand, would appear to take the position that objectives and outcomes can be mandated through the curriculum and that they can consequently be the same for all students. Assessment theorists such as Sadler (1989) take a middle position and argue that while objectives and outcomes can be defined at a level beyond the student, they need to be made explicit and transferable such that students come to possess a knowledge of quality and performance that is in accordance with the globally defined objectives and outcomes.
In comparing and contrasting discussions of the term feedback in science and in education, it has already been noted that whereas 'the system' as it is understood in science can be easily defined - if not always controlled - there is no such simplistic definition for the term within the educative context. It could well be that this is why the term appears infrequently in discussions of educative feedback. Potentially, however, a full interrogation of what exactly constitutes an educative system or systems and the impacts of these on teaching, learning and assessment, could yield valuable insights into the ways in which educational feedback can and should operate, particularly with respect to how effective feedback comes to be utilised at the level of teacher-student interaction.
To return to the example of the heating system outlined earlier, if a thermostat has been programmed to respond to changes in thermal differentials and calibrated according to a set temperature but is either; (a) unable to measure the temperature according to this differential or, (b) unable to communicate the information regarding increases or decreases in temperature to the heater, than in either case the provision of the thermostat would be redundant as the heater still would not work. In a heating system the thermostat itself through the use of a bimetallic strip is responsible for ensuring that temperature is measured, compared to the set temperature required and that this information is fed back to the heater in an appropriate form such that the heater is turned on or off.
The function of the thermostat and the bimetallic strip cannot be simplistically transferred to an educative context. Nonetheless, it must be noted, that in many instances teachers are currently responsible for ensuring that student achievement is measured and compared to defined objectives and anticipated outcomes, and are simultaneously responsible for ensuring that this information is fed back to students with the aim of assisting them to improve performance. At the moment then it would seem that teachers have unproblematically been given the responsibility of acting in a similar manner to a mechanical device. As a consequence, the complex issues of providing effective feedback to students within the area of classroom teaching, learning and assessment have not been fully or systematically researched. It is therefore timely and indeed necessary for sustained research into what may be described as the pedagogy of feedback to be undertaken.
Reference List
Arthur, H. (1995). Student self-evaluations: How useful? How valid? International Journal of Nursing Studies. 32(3). pp 271-276.
Bangert-Drowns., R.L., Kulik, J.A., & Kulik, C.C. (1991). Research Effects of Frequent Classroom Testing. The Journal of Educational 85(2) pp. 89-99.
Bernard, R. & Naidu, S. (1992). Post-questioning, concept-mapping and feedback: a distance education field experience. British Journal of Educational Technology. 23(1). pp 48-60.
Butler, D. & Winne, P. (1995). Feedback and Self-Regulated Learning: A Theoretical Synthesis. Review of Educational Research. 65(3). pp. 245-281.
Daws, N. & Singh, B. (1996). Formative Assessment: To what extent is its potential to enhance pupils' science being realised. Science Education Notes. 77(281). pp93-100.
Gipps, C. (1994). Beyond Testing: Towards a Theory of Educational Assessment. London. Falmer Press.
Elshout-Mohr, M. (1994). Feedback in Self-Instruction. European Education. 26(2). pp. 58-73.
Hughes, H.S. (1958). Consciousness and Society. in P Worsley (Eds) (1985). Modern Sociology. Middlesex. Penguin Books.
McCurdy, B. & Shapiro, E. (1992). A Comparison of Teacher, Peer and Self Monitoring within Curriculum-Based Measurement in Reading Among Students with Learning Disabilities. The Journal of Special Education. 26(2) pp 162-180
McMeniman, M. (1986). Formative and Summative Assessment: A Complementary Approach. Assessment Unit Discussion Paper No 6. Brisbane. Board of Senior Secondary School Studies.
Mory, E.H. (1992). The Use of Informational Feedback in Instruction: Implications for Future Research. Educational Technology Research and Development. 40(3). pp 5-20.
Ramaprasad, A. (1983). On the definition of feedback. Behavioural Science. 28. pp 4-13.
Sadler, R. (1989). Formative Assessment and the Design of Instructional Systems. Instructional Science. 18(1). pp. 1-25.
Scientific Apparatus Makers Association. (1970). Process Measurement and Control Terminology. New York. Process Measurement and Control Section.
Slavin, R.E. (1987). Mastery Learning Reconsidered. Review of Educational Research. 57(2). pp255-271
Conference Entry Page About the Conference Abstracts Conference Program
TEDI Home About Assessment UQ Home
This site built on a Macintosh using Dreamweaver.. Last modified 20/1/99; 10:14:16 AM.