A Pedagogy for Learning Objects: The Theoretical and Empirical Basis

Wallace Hannum
University of North Carolina at Chapel Hill USA

Key Words: learning objects, learning theory, empirically-based principles Abstract:

This paper discusses the differences between learning objects that simply present content to learners and learning objects from which students do indeed learn. This difference is found within the theoretical and research literature on learning and instruction. This paper examines theoretical concepts from cognitive and constructivist learning theory in the context of learning objects to create an understanding of how learning objects can support mLearning and to distill several theoretically-based principles for the design and use of learning objects in mLearning. This paper also examines the research literature to identify instructional variables that influence learning outcomes to identify those variables that have been shown to produce superior learning which can be applied in mLearning. All learning objects are not equal in their ability to produce learning. Learning objects that are built to embody certain instructional conditions and events within mLearning environments will produce greater learning outcomes than learning objects that simply present pieces of information.

Introduction

If learning objects are to achieve even a portion of their potential to influence how we develop and deliver instruction, we must focus on ensuring that learning objects do indeed produce learning. To date considerable effort has centered on issues of standards for metadata and interoperability of learning objects (SCORM, IEEE, IMS, Dublin Core, ARIADNE). As Bannan-Ritland, et al. (2002) noted, most of the effort has focused on technological attributes, metadata standards and system specifications. Such efforts are essential in reaching the goals of development and widespread use of learning objects. However, if the technical issues associated with learning objects obscure the pedagogical issues we are likely to have a period in which the use of learning objects becomes more widespread, but that will be followed by a period of abandonment when the use of learning objects fails to contribute substantially to learning outcomes. 1.1 Emphasis on outcomes

We are in a time when more emphasis in K-12 education has been placed on the learning outcomes of students. Certainly educators are likely to resist or abandon the use of learning objects if they fail to see an increase in students achievement as a

result of using learning objects. Similarly in higher education, faculty are likely to embrace the use of learning objects only when they see that their students learn more when they incorporate learning objects into their classes. Corporate education has always placed an emphasis on results, so they will be likely to use learning objects only if they see a direct benefit in terms of better learning outcomes (effectiveness) or the same amount of achievement in less time (efficiency). While there are considerable differences in education at these different levels, all share an emphasis on promoting student achievement. This is the backdrop against which they judge the adequacy of learning objects. Simply put, if the use of learning objects fails to produce measurable improvements in student outcomes, then most educators and trainers will decline to make much use of learning objects despite the appeal of the technology. 1.2 Diffusion of learning objects

While some educators and trainers have an intense interest in learning objects, they likely represent what Rogers calls the early adopters, not the majority (Rogers, 1995). The very nature of learning objects requires that to be successful, more than just the early adopters begin to use learning objects. This follows from the economics of learning objects. Learning objects do not have value as a result of their scarcity in the way that diamonds are valuable because they are rare. Rather learning objects become valuable for the opposite reasonwhen they are plentiful. This is like the economics of fax machinestheir value increased as they became more plentiful, not when they were scarce. If learning objects are to have more than a trivial impact on education and training, we must have many people contributing and using learning objects. Thus, we must get beyond those few early adopters of learning objects. This requires that learning objects are shown to be effective in terms of producing higher achievement. Simply put, showing that learning objects are technologically feasible is not sufficient to the majority of potential users of learning objects. Such a scenario is likely developing as a result of our lack of attention to matters of an appropriate pedagogy for learning objects. In essence, we are finding ways to create repositories of learning objects that meet technical requirements for being selfcontained, reusable, small units of content that can be easily shared and aggregated. However, when learners complete lessons formed from these learning objects they may neither learn more nor learn faster than before. Our work on learning objects will be a technical success and an instructional failure without attention to theoretical and empirical work on learning necessary to establish requirements for producing learning. 1.3 Purpose.

This paper addresses concerns associated with learning objects by examining: 1) the theoretical basis of learning relevant to the design and use of learning objects, and 2) the empirical literature on the effectiveness of different instructional variables that can inform the creation of learning objects. The paper then offers recommendations for the design and use of learning objects based on this evidence about learning.

Information and Learning

There is little magic in learning objects. Simply putting some piece of content into a learning object repository does not ensure that a person using that learning object will in fact learn from that use. In short, the presentation of content does not equate with learning. This distinction is fundamental to the development of effective learning objects. While content is necessary for instruction and learning to take place, presenting content alone is far from sufficient (Stolovitch & Keeps, 2002). Large amounts of content are presented to students daily in schools either directly by teachers or through books or via technology. Certainly students hear and view considerable amounts of content during a school day. Yet most of that content is not learned to a level of mastery by most of the students as a result of that exposure. The research on learning through lecture is informative on this point. Students typically acquire only a portion of the content they are exposed to through lectures, and they experience a sharp decline in retention of that content within a few days. Undoubtedly, students learn from lectures, but research and our own experience would lead us to conclude that presenting a lecture to students on some content does not ensure their mastery of that content. A similar situation arises if we give students content in the form of written passages for them to read. It is likely that giving students content in the form of learning objects will not guarantee success in terms of their learning. The central question is why is some content that is intended to instruct -- whether in the form of a lecture, a written passage, a video, or some other form -sufficient to cause learning to happen while much other content in the same form that also is intended to instruct fails to cause learning to occur? That is, under what conditions will learning occur when we're presenting content for the purpose of instructing?

3 Learning Theory
3.1 Cognitive theory

During the last half-century, many people -- notably educational psychologists -- have attempted to identify those conditions or circumstances that when present are most likely to result in students learning instructional content. Gagne (1985) sought to identify those imported instructional conditions that gave rise to different types of learning outcomes. According to Gagne, the learning of intellectual skills -- such as the ability to solve problems, apply rules, and use concepts -- depends on the prior mastery of and ability to recall the specific prerequisite skills (Gagne, 1985). The learning of information -- such as names, labels, facts, or larger organized bodies of information -- depends on the information being presented in a clear, organized fashion that stresses interrelationships with other information, presenting information in a context that is meaningful to the learner, the opportunity to practice recalling this information, and providing the learner with feedback on the correctness of his/her response when they practice recalling information. The learning of motor skills depends on the learner observing a model of the correct execution of motor skill, having the opportunity to practice the motor skill, and receiving corrective feedback that would allow him/her to modify their execution of the motor skill such that it more closely approximates the correct execution of the motor skill. If we consider developing learning objects within Gagne's framework then we would develop different types of learning objects depending on the type of content that we were teaching in order to ensure successful learning. It we wanted to teach an intellectual skill, such as the ability to solve a problem, our learning objects should be created to 3

stimulate recall of the prerequisite skills that underlie the ability to solve that specific problem. If we wanted to teach information, such as the ability to describe the causes of World War II, our learning objects should be created to help the students: 1) see how this content is organized and relates to other content, 2) form links between this new content and content that they already know, 3) understand how this content can be meaningful to them, 4) practice storing and retrieving this content from their memory, and 5) get feedback on the correctness of their practice. Learning objects they conform to this are likely to be successful based on Gagne's theory, and learning objects that fail to take these learning conditions into account are less likely to be successful. Another aspect of Gagne's theory of learning that has relevance to the design of learning objects is his concept of instructional events. According to Gagne et al. (2005) learning is best understood within information processing framework that views learning as result of the way humans process information. Based on this conception of learning as the processing of information, Gagne created a model for how instruction should take place such that it maximizes the probability of the internal processing of information being successful. In short, Gagne identified those specific things that instruction should do so that it better fits with how people learn. Gagne referred to this model as the instructional events model and indicated that these were the steps through which a lesson should progress to ensure that students will learn from the lesson. In Gagne's view, just presenting instructional content to learner would not be sufficient. Rather instructional content should be organized into nine instructional events and sequenced in a particular fashion for learning to be successful. Gagne's nine instructional events are: 1. Gain the learner's attention 2. inform the learner the objective 3. stimulate recall of prerequisites 4. present the new learning material 5. provide learning guidance 6. provide opportunity for practice 7. provide feedback 8. assess the learning effectiveness 9. promote transfer and retention of learning Gagnes instructional event model has direct implications for the design of learning objects. This model suggests that not all learning objects will be successful, but that different types of learning objects that correspond to the different instructional events are required at different points in a lesson. To implement this model, we should have learning objects that gain the learner's attention, that inform the learner of the objectives, that stimulate recall of prerequisites, and so forth. People who are using learning object repositories to create lessons would have to know for each learning object where it fits within the instructional event model. Thus a specific learning object would have to be tagged according to which instructional event it activates. When planning a lesson using this model a person would query a learning object repository according to which instructional event it followed. By no means is Gagne's theoretical framework the only one available to help people ensure that the learning objects we create and use will cause students to learn the content we're trying to teach. Another framework for examining learning outcomes

and the conditions necessary to produce these outcomes has been developed by Merrill (2001, 2002). Merrill contends that many instructional products fail to effectively teach. This confirms the view that simply providing content is not sufficient to cause learning. Merrill (2002) put forth a set of first principles of instruction that he indicated were consistent with most instructional design models. In his view, instruction would be more successful when it was consistent with these first principles. The following is taken directly from Merrill (2002, p17). First Principles of Instruction Demonstration principle Learning is promoted when learners observe a demonstration of the skills to be learned that is consistent with the type of content being taught. Demonstrations are enhanced when learners receive guidance that relates instances to generalities. Demonstrations are enhanced when learners observe media that is relevant to the content. Application principle Learning is promoted when learners engage in application of their newly acquired knowledge or skill that is consistent with the type of content being taught. Application is effective only when learners receive intrinsic or corrective feedback. Application is enhanced when learners are coached and when this coaching is gradually withdrawn for each subsequent task. Task-centered approach Learning is promoted when learners are engaged in a task-centered approach which includes demonstration and application of component skills. A task-centered approach is enhanced when learners undertake a progression of whole tasks. Activation principle Learning is promoted when learners activate relevant cognitive structures by being directed to recall, describe or demonstrate relevant prior knowledge or experience. Activation is enhanced when learners recall or acquire a structure for organizing the new knowledge. Integration principle Learning is promoted when learners integrate their new knowledge into their everyday life by being directed to reflect-on, discuss, or defend their new knowledge or skill. Integration is enhanced when learners create, invent, or extrapolate personal ways to use their new knowledge or skill to situations in their world. Integration is enhanced when learners publicly demonstrate their new knowledge or skill. The concept of first principles gives us insight into what kinds of learning objects are likely to be effective. For example the demonstration principle suggests three different kinds of learning objects: 1) a demonstration of the skill to be learned, 2) 5

specific guidance while learning the skill, and 3) media relevant to the content. The third kind of learning object suggested by this principle media content is also consistent with the observational learning theory of Bandura who holds that we learn from observing a model execute a performance. The first principles of Merrill can be the basis for determining what kinds of learning objects are likely to be effective. This provides a basis for selecting learning objects such that we dont simply create content that has little or no instructional value. The models of Gagne and Merrill are cognitive models of learning. While there are some differences between these models, their similarities are much greater. Both assume there is a structure to knowledge that is external to the learners. Both believe that there are different kinds of learning outcomes, and that we should pay attention to these differences when planning instruction. Both believe that there are different ways of providing instruction for different categories of learning outcomes. 3.2 Constructivist theory

A different point of view about learning is found in constructivist theory of individuals such as the Vygotsky, Bruner, and Jonassen. Constructivists believe that the content of what is to be learned is not present to the learner in its final form, but rather the learner must construct meaning from the content for himself or herself in order for the learning to be worthwhile. Learning objects developed in this perspective would be less prescriptive than those developed within a perspective similar to Gagne or Merrill. Constructivist theory would allow students to have more direction and control over selection and sequencing of instructional content. They would be given more options over what content to explore perhaps with some suggestions but not with prescriptive control over the sequencing of instructional content. Constructivists would have variety in their learning objects but these would differ from anything like Gagnes instructional events. Constructivists caution against learning objects that are based on the work of Gagne and Merrill because the constructivists have a fundamentally different view of learning and instruction. The assumption that we can apply certain pedagogical rules to prescribe the specific learning objects from which a specific learner will learn at a specific moment is rejected by constructivist theories although this position is consistent with cognitive theories. The constructivist position is that we simply do not have sufficiently robust learner models to predict which learning objects will work in a specific situation. They would rely on the learners to make their own choices. Learners may elect to view instructional content in different ways and would find having only one way prescribed for them to view content at any point in time to be limiting. There is also the question of the instructional content itself. As Jonassen (1999) has noted, constructivists view knowledge as often being ill-structured. This argues against the highly prescriptive approaches to learning objects that have rigidlystructured learning objects that are not particularly flexible. Perhaps the greatest distinction between constructivist theory and cognitive theory is their view of the learning process itself. Constructivist theory sees learning as happening through activity on part of the learners within a context. Learners create or co-create meaning while interacting within context. On the other hand, learning within cognitive theory

is more a reception of knowledge from carefully structured content than it is an exploration of knowledge from ill-structured content. Based on this distinction, these theories would have different kinds of learning objects and would use them in different ways. 3.3 Views concerning learning objects

The different views of learning between cognitive theory and constructivist theory influence how each theory would approach the development and use of a learning objects repository. The idea of learning objects as small reusable pieces of content would have appeal to cognitive theory if the learning object identified both the type of learning outcome and the instructional event/transaction to which the learning object referred. Only then would learning objects be valuable, since when planning instruction we need to know the type of learning outcome we seek, and we need to include the relevant instructional conditions that will cause this learning outcome to be met. From the cognitive point of view the content in any learning object repository must be tagged according to the type of learning outcome and the type of instructional event or factor. For example, it would not be sufficient to tag a learning object with information about its name, source, content area, size, intended learner, suggested learning time, type of resource, or mode of use. Rather, cognitive theory would suggest that a learning object repository would include information about the type of learning and the specific instructional event associated with each learning object. Learning objects within a constructivist position would be much less prescriptive and much less structured. Their learning objects would be flexible and not crafted according to a set of principles like Gagnes instructional events or Merrills first principles. They would seek to put learners in a context and involve them in activities (Jonassen, et al., 2004). This is more a guided exploration of content within a context with considerable flexibility for the learners. This is not the same as giving learners access to the Internet and encouraging them to go learn something! A constructivist learning environment is planned and structuredjust not in the same fashion as in cognitive theory. Ip and Morrison (2001) demonstrated how different pedagogical conceptions could result in different kinds of learning objects. At one end of their continuum of pedagogy the learning objects would be drill and practice items. Moving towards the other end of this continuum the learning objects would be simulations, cognitive tools such as databases, spreadsheets and semantic networks, and indexed resources. This matches the shift from cognitive theory to constructivist theory and shows how learning objects would differ based on our assumption about learning and instruction. A good place to start when considering learning objects is with theoretical concepts of learning and instruction. As briefly noted above, different theoretical positions would guide us towards different kinds of learning objects. Our views about learning shape our concept of learning objects.

Research on Learning and Technology

Considerable research over the past four decades has investigated the impact of technology on learning. While a thorough review of this literature is beyond the scope of this paper, several key points from this literature have relevance to the discussion 7

of learning objects. When technology has been used for instructional purposes the resulting impact on learning outcomes has been marginal. Despite great proclamations of the value of using technology for instruction, the empirical research urges much more caution. A common finding from the research that compares technology-based instruction with traditional instruction is that of no significant differences. This has created the debate not only about whether technology does improve instruction but also whether technology ever could improve instruction. The opposing points of view are stated most succinctly by Clark (1983, 1994), who indicated that technology is merely a means for delivering instruction and doesn't improve the quality of instruction, and Kozma (1994), who indicated that when used correctly technology does improve the quality of instruction. Considerable, but perhaps not definitive, empirical evidence addresses this issue of whether the use of technology improves the quality of instruction. Table 1 shows some effect sizes from meta-analysis of the effects of technology on student learning outcomes. Note that these effect sizes are small to moderate with the largest still less than a half a standard deviation. Table 1. Effect Sizes for Technology Technology Application Multimedia instruction Integrated Learning Systems -- basic skills CAI elementary CAI secondary Integrated Learning Systems mathematics CAI -- reading CAI college Integrated Learning Systems reading Distance learning Interactive distance learning Effect Size 0.40 0.38 0.30 0.187 0.18 0.13 0.127 0.06 -0.0128 -0.028

As Table 1 shows, the use of technology typically produces small to moderate improvements in learning outcomes. Contrast this with the effect sizes for a variety of instructional variables shown in Table 2. As you note in this table, many instructional variables produce substantially higher levels of learner achievement than does the use of technology. In fact, all these instructional variables resulted in greater student achievement than did any of the uses of technology. Perhaps learning objects will show a larger effect than technology use on education has shown before, but this is doubtful. As Hannum (2007) noted, it is not the use of technology itself that produces learning gains but rather the use of empirically-based instructional practices that can be delivered by technology. Table 2. Effect Sizes for Different Instructional Variables Variable One-to-One Tutoring Analogies Comparison Graphic Representations 2.0 1.65 1.32 1.24 8 Effect Size

Generative Activities Summarizing Note-taking Corrective Feedback Direct Instruction Questioning Reciprocal Teaching Homework Cooperative Learning Mastery Learning Frequent Testing Concept Mapping Advance Organizers

1.14 1.0 0.99 0.94 0.93 0.93 0.86 0.77 0.59 0.53 0.48 0.45 0.44

There seems little doubt that the addition of technology alone to the learning process is not likely to produce significant learning gains. In a review of twenty years of research on technology in education Maddox (2004) stated that article after article and report after report have concluded that the effect of information technology on teaching and learning has been disappointingly minimal. It is still early to estimate the effects on student achievement of using learning objects, but if the pattern holds it is likely that using learning objects might make technology-based instruction easier, quicker, and less expensive to develop but not more effective in producing student learning. If we desire to produce learning gains we should look to other factors than technology itself.

Implications

Most likely all learning objects are not equal in their ability to produce learning. Unless learning objects are built to embody certain instructional principles that have been shown to produce greater learning outcomes, they are not likely to add value in terms of student achievement. Simply placing some content into a learning object repository is poor pedagogy that will likely produce marginal learning gains at best. Basing the choice of learning objects on instructional variables that have demonstrated effectiveness will result in greater student achievement. Following the empirical evidence regarding instruction, we would be wise to create learning objects that embody analogies, comparisons, graphic representations, corrective feedback, concept mapping, and advanced organizers to name a few of those variables that have shown to be effective in producing student learning. While there are theoretical differences concerning learning that continue and have implications for the design and use of learning objects, this does not suggest that we fail to pay attention to theory. Both cognitive and constructivist theory offer guidance in how to produce learning. Unfortunately, many learning objects are based on neither constructivist or cognitive theory but remain atheoretical items of content. Without a basis in theory and empirical evidence, learning objects are little more than simple pieces of information that may, or more likely, may not support learning. The technology for these learning objects may work but the learning will be absent. We should take seriously the word learning in learning objects and base more of our work on theory and research concerning learning.

References:
[1] Bannan-Ritland, B., Dabbagh, N., & Murphy, K. (2002). Learning object systems as constructivist learning environments: Related assumptions, theories and applications. In D.A. Wiley (Ed.) The instructional use of learning objects. Bloomington, IN: Agency for instructional technology and association for educational communications and technology. [2] Clark, R.E. (1983). Reconsidering research on learning from media. Review of Educational Research, 53, 445-459. [3] Clark, R.E. (1994). Media will never influence learning. Educational Technology Research and Development, 42(2), 21-29. [4] Gagne, R. (1985). The Conditions of Learning (4th ed.). New York: Holt, Rinehart & Winston . [5] Gagne, Robert M., Wager, Walter W., Golas, Katharine, and Keller, John M. (2005). Principles of Instructional Design 5th Ed. Belmont, CA: Thomson/Wadsworth. [6] Hannum, W.H. (2007). When computers teach: A review of the instructional effectiveness of computers. Educational Technology, 47(2), p5-13. [7] Ip, Albert and Morrison, Iain (2001). Learning objects in different pedagogical paradigms. Proceedings of Ascilite. 2001. [8] Jonassen, D. (1999). Designing constructivist learning environments. In C. M. Reigeluth (Ed.), Instructional design theories and models: A new paradigm of instructional technology (Vol. II, pp. 215-239). Mahwah, NJ: Lawrence Erlbaum Associates. [9] Jonassen, D.H. (2004). Problem solving: The enterprise. In J.M. Spector, C. Ohrazda, D. Wiley & A. Van Schaak (Eds.), Innovations in instructional technology: Essays in honor of M. David Merrill (pp. 91-110). Mahwah, NJ: Lawrence Erlbaum Associates. [10] Kozma, R.B. (1994). Will media influence learning? Reframing the debate. Educational Technology Research and Development, 42(2), 7-19. [11] Maddux, C.D. (2004). The web in k-12 education: Is there a future? Computers in the Schools, 21(3/4), 149-165. [12] Merrill, M. D. (2001). Toward a theoretical tool for instructional design. Instructional Science, 29(4-5), 291-310.

10

[13] Merrill, M. D. (2002). Knowledge objects and mental models. In D. A. Wiley (Ed.), The Instructional Use of Learning Objects (pp. 261-280). Washington DC: Agency for Instructional Technology & Association for Educational Communications and Technology. [14] Rogers, Everett (1995). The Diffusion of Innovations 4th Ed. New York: The Free Press. [15] Stolovitch, H. D., & Keeps, E. K. (2002). Telling Ain't Training. Alexandria, VA: American Society for Training & Development Press.

Author:
Wallace Hannum Associate Professor and Program Coordinator University of North Carolina at Chapel Hill, Educational Psychology, Measurement and Evaluation Peabody Hall CB#3500 Chapel Hill, NC 27599-3500 USA Hannum@unc.edu

11