Component Display Theory (CDT) provides a list of prescriptions for designing instruction for different kinds of instructional outcomes. As we moved toward trying to automate the instructional design process we found that CDT was not precise enough to allow computer implementation of expert system technology which will prescribe instruction. Instructional Transaction Theory (ITT) is an attempt to provide more precision to CDT thereby making automated instructional design a possibility. This increased precision also has value for instructional designers in that it provides a more precise way to describe knowledge representation, instructional strategies, and instructional design prescriptions.
Component Display Theory was an attempt to identify the components from which instructional strategies could be constructed. CDT describes instructional strategy in terms of strategy components: primary presentation forms (PPFs), secondary presentation forms (SFPs), and interdisplay relationships (IDRs). CDT identifies strategy prescriptions for different kinds of learning outcomes. Each of these prescriptions identified a best case combination of PPFs, SPFs, and IDRs for a particular kind of learning outcome. CDT was analysis oriented, emphasizing the components of instructional strategies for different kinds of instructional goals.
ITT is synthesis oriented, emphasizing the integration of these components into instructional transactions. An instructional transaction is all of the interactions necessary for a student to acquire a particular kind of knowledge or skill.
The presentation of ITT in this chapter emphasizes what Reigeluth calls component methods. In this paper we will introduce a methodology for representing knowledge in the form of knowledge objects and elements (slots) of knowledge objects. Knowledge objects and their elements provide the components of subject matter content (knowledge). ITT describes instructional strategy as methods (algorithms) for manipulating the elements of knowledge objects.
Instructional theory is concerned with two primary considerations: What to teach? And how to teach?
What to teach has two considerations: selection and representation. ITT is not concerned with the curriculum selection question of what should be taught, but rather having selected what should be taught, what are the knowledge components required for a given type of instruction? And how should these knowledge components be represented to facilitate instructional design?
How to teach specifies the way that these knowledge components are presented to the student in order to engage the student in an interaction which is appropriate for promoting the acquisition of the knowledge or skill that is the goal of the instruction. Instructional strategies include the presentation of the appropriate knowledge components, practice with or student activities involving these knowledge components, and learner guidance to facilitate the student's appropriate interaction with these knowledge components.
Authoring systems for CBT are based on a database model of computing. The student is presented a record containing subject matter content (a program frame). The program then presents the student with one of several options: press a key to see the next record; select an item from a menu to see the next record; or respond to a question and the next record will be determined based on the answer given. This model of instructional computing has one serious limitation: except for the branching strategy, all other instructional strategies are hidden in the record (frame) and therefore transparent (not available) to the instructional system for additional processing. The instructional strategies to be used must be determined by the designer of the system and incorporated within the records of the data base.
Outside of instruction many computer programs are based on an algorithmic model. In this model data is manipulated by one or more sets of instructions for processing (displaying, transforming) this data. If the knowledge to be taught is thought to be data, and the strategies for teaching this knowledge are thought to be instructional algorithms, then an algorithmic model of computing can also be applied to instruction. However, an algorithmic instructional system requires that the knowledge be accessible in a form that lends itself to processing by the instructional algorithms. A primary focus of ITT is to describe such a knowledge representation system.
ITT is an algorithmic instructional system. Knowledge is represented as data. The components of this knowledge are processed (displayed, transformed) by the instructional algorithms built into an instructional system. While this representation was specifically designed to facilitate the design of computer-based instruction, this form of representation also has value for designing instruction for other modes of delivery.
Effective instruction. First, we are concerned with the current emphasis on information and the lack of emphasis on appropriate instructional strategies. By describing instructional strategies as algorithms (transactions) for manipulating data structures (knowledge objects) we have provided a much more precise description of the different kinds of instructional transactions required for different kinds of instructional outcomes (goals or objectives). Our hope is that this formulation will enable instructional designers to design more effective and appealing instructional products. Furthermore, by building these transactions into instructional development tools there is an increased probability that the resulting instructional interactions will be based on sound principles of instructional design.
Efficient instructional development. Second, our intent was to derive theory and methodology which would facilitate the automation of much of the instructional design process. Instructional development is a labor intensive industry. If we are to obtain efficiencies in the development of large amounts of computer-based, interactive, multimedia instruction then we must significantly increase our design and development efficiency. We believe that building appropriate instructional transactions into instructional development tools will enable automating portions of the instructional design process and will enable us to realize this efficiency.
Instructional learning environments. Third, interactive learning environments (instructional simulations and microworlds) are extremely labor intensive and hence very expensive to develop using existing technologies. Representing knowledge as knowledge objects enables the building of a general purpose simulation engine. This makes possible a learning environment builder which enables the efficient development of these more effective instructional interactions. Furthermore, since tutorial instruction and learning environments are based on the same knowledge representation this architecture enables effective learner guidance to overlay instructional learning environments.
Adaptive instruction. Fourth, ambiguous representation of knowledge and imprecise specification of instructional strategies has hindered the development of truly adaptive instruction. The precise representation of knowledge in the form of knowledge objects and the representation of instructional transactions as algorithms for manipulating this knowledge makes possible instructional strategies that can be adapted to an individual learner in real time as they interact with the instructional materials.
The component methods of ITT can be used to describe any instructional strategy whether tutorial or experiential. However, in this presentation we will emphasize only a few of the possible instructional goals to which ITT can apply. The example presented for illustration has as a primary goal learning a procedure. In support of this goal the learning environment includes guidance for learning the function and location of the parts of a device. Also in support of the procedural goal the learning environment includes guidance for learning to predict consequences and trouble-shoot unexpected consequences by identifying the underlying conditions which must be satisfied before a given consequence can occur.
Figure 1 illustrates a learning environment designed to teach the learner how to install or remove a double seat valve. This instruction was prepared to train technicians who must maintain valves in dairies and breweries.
Prior to entering the learning environment the learner is given the following goal:
This work task involves properly disconnecting all hoses and wires from the valve and correctly unbolting and removing the valve insert from the pipeline.The learning environment consists of a diagram of the valve showing all of the hoses and wires connected to the valve. The learning environment is an "open-ended learning environment 3". Clicking on a given part of the diagram brings up an action menu. For example: clicking on the switch brings up a menu with the action flip. Selecting this action causes the switch to change positions (down to up) and produces an audible click. This action also has the consequence of setting a property of the compressor to off. Clicking on the connector for the air hose brings up the action undo which when selected disconnects the air hose from the valve. Hoses and wires can be connected and disconnected at will except as constrained by the conditions of the system. An example of a constraint is that you cannot disconnect the air hose until the compressor is off. By exploring the system most learners can eventually figure out how to disconnect the valve and remove it from the pipe.
The learning environment is supported by learner guidance of various types. This learner guidance implements various different instructional strategies designed to teach the student knowledge about the valve and the skill of removing the valve from the pipe.
Clicking on the right button of the mouse causes a functional description to appear in a pop-up window near the part under the cursor. For example, right clicking on the air connection brings up the illustration and scrolling description illustrated in Figure 2. Note that the functional description in this system could be any media including graphic, video, audio, text or a combination. The authors chose to use a picture and text caption. The student can thus "explore" the function of each part of the device.
Clicking on the Guide button on the control panel at the bottom of the screen pops up a menu of guidance options. One option is, Tell me about some of the parts. Selecting this guidance causes a "lecture" to be provided about the parts of the system. The guide presents the same type of information as shown in Figure 2 for each part in turn. The student has control over the pace of the lecture by indicating when they are ready for the next part to be presented.
Another guide option is, Let me locate the parts. The guide then presents a name of a part and the student is required to point to (click when the cursor is over) the part. Another guide option is, Let me name the parts. The guide highlights a part and the learner is required to select the correct name from a list. Another guide option is, Let me identify the functions. The guide presents a description of a part and the student must click on the corresponding part. Right-wrong-correct-answer feedback is provided after each response. These practice activities incorporate sampling with replacement such that when a student misses an item it is put back into the list and presented to the student again until the student has correctly responded to each part. At the end of a practice exercise the student is given their score indicating how many tries were required to correctly name, locate, or identify the function of each part.
One of the goals of ITT is to enable an instructional development system to automatically generate appropriate instruction. In the Instructional Simulator the instructional strategies (presentation and practice) are built-into the system. The designer merely describes the knowledge objects and the system automatically generates these presentation and practice strategies.
Given enough time most students can eventually "discover" the procedure from removing the valve from the pipe. However, "raw discover" is inefficient and often results in a trial-and-error approach to the performance in the real world. Procedure learning is more efficient when appropriate learner guidance is provided. The goal of procedure learning in this learning environment is to learn the necessary and sufficient steps for removing and replacing the valve in the pipe.
The guide provides four levels of procedure practice. Level one is a "hands-off" demonstration. In this demonstration the guide performs each of the steps in turn. The cursor moves to the appropriate part, the appropriate action is indicated, the guide then performs the appropriate action, and the system illustrates the consequence of this action. Level two is a "Simon Says" demonstration/simulation. In this demonstration the guide tells the student the step to perform. For example, "Flip the air compressor switch." If the student attempts to do any other action the guide presents a message such as, "That is not the air compressor switch." Thus the student can only perform the step requested by the guide. After the student has selected the action, the system illustrates the consequence of that action.
Level three is a "Do the next step" simulation. In this practice the guide presents the message, "Do the next step." If the student attempts to do different step the guide presents a hint such as "That is not the air compressor switch," After the student has selected the action, the system illustrates the consequence of that action.
Level four is "performance" or "You-do-it" simulation. In this practice the student can perform any of the steps in the procedure and see the consequence of this step subject to the constraints of the system (that is, some steps cannot be performed until a prior step has been completed. For example, you cannot remove the valve until you have removed the flange bolts, inserted a bolt in the tap hole to break the seal, and removed the bolt from the tap hole). When the student believes they have completed the procedure they click a finished button. The guide then shows the steps required for the shortest path to the goal and also shows the steps taken by the student. Unnecessary or incorrect steps in the students path are highlighted in red. The number of steps required by the student to accomplish the goal are recorded in the student record.
ITT knowledge objects enable all of these levels of practice to be built-into the system. The designer merely provides the elements of the knowledge objects required by the simulation and the system automatically creates the various levels of practice available in the system. The designer can select some or all of these levels of practice for a given learning environment.
It is one thing to learn the steps in a performance and to be able to carry them out in order to accomplish some goal (such as removing a valve from a pipe). However, when learners have an explanation of each step in the procedure that identifies what happened and why it happened, their ability to retain the procedural skill is enhanced. In addition, knowing what happened and why is necessary to problem solving or trouble-shooting a device or system. What happened indicates the consequence of a given action. Often what happened can be observed by the student in changes in the appearance of the system. In some cases what happened may change a condition of the system which does not show up in the physical appearance of the system. Why indicates the conditions which must be met for a given consequence to occur. When a student performs an action and nothing happens or something unexpected happens, then an explanation indicates what conditions were not met or what conditions led to the unexpected consequence. The goal of explanation is to enable the learner to "predict" what will happen under specified conditions, or to "explain" (identify the conditions which were not met often called trouble-shooting) when a consequence fails to occur or when an unexpected consequence occurs.
The guide provides three levels of explanation. One guidance option is Explain. During free exploration of the system the student can request an explanation after any action. The guide presents a description of what just happened and why it happened. For example, the student attempts to remove the valve. Nothing happens. The student requests an explanation. The guide provides the following message: "When you attempt to remove the valve from the pipe nothing happens. This is because the flange bolt is still in the tap hole." The Explain function can also be turned on during any of the practice levels. The explain display is updated after each action by the student.
A second guidance option is Predict. The guide configures the system and asks the student to select from a list "What happens next?" and "Why?" The student can then confirm their prediction by executing the next step(s) in the procedure and observing what happens. The students accuracy in prediction is recorded by the system.
A third guidance option is Trouble-Shoot. The guide configures the system, sometimes introducing a fault. The student is requested to carry out the next step(s) in the procedure and to explain "What happened" and "Why." In this situation the student provides the explanation rather than the guide. The student selects what happened and why from a list of consequences and conditions.
ITT knowledge objects makes it possible for the designer of a learning environment which includes an explanation system to instantiate (provide information for) the elements of the knowledge objects required for the simulation and the system will automatically generate the various levels of explanation.
Robert M. Gagné (1965, 1985) stated as a primary assumption of instructional theory that there are different kinds of learning outcomes (learning goals) and that each of these different kinds of learning outcomes required unique conditions for learning. An appropriate instructional strategy incorporates all of the necessary conditions for presenting the knowledge or demonstrating the skill, providing practice, and providing learner guidance for a given type of learning outcome. Gagné indicated nine events of instruction which include these three phases on instruction. Appropriate conditions for learning always require all of these activities. Information which does not include presentation, practice, and learner guidance is information but not instruction. Different instructional outcomes (objectives) require different types of presentation, different types of practice, and different kinds of learner guidance. It is this difference in the required conditions for learning that distinguishes different kinds of learning outcomes.
In previous papers we have called an instructional strategy that incorporates all of the conditions for teaching a given type of learning an instructional transaction. We previously identified 13 classes of instructional transactions (Merrill, Jones & Li, 1992). In this paper we shall describe only three of these classes in terms of instructional algorithms for processing the elements of knowledge objects. These instructional transactions include: IDENTIFY (component or naming or parts of E', also related to facts in CDT ); EXECUTE (activity or procedures or how to E', procedures in CDT), and INTERPRET (process or what happens E', principles in CDT). The description of the other transactions identified in Merrill, Jones & Li (1992) will have to wait for another paper at some future time.
In ITT an instructional transaction is all of the learning interactions necessary for a student to acquire a particular kind of knowledge or skill (learning goal). The instructional algorithm (called an instructional transaction shell) required to promote an appropriate instructional transaction operates on a set of knowledge objects, that are related in a particular way (knowledge structure), that contain all of the knowledge that is required in order for a student to acquire the instructional goal. An instructional transaction algorithm includes the presentation strategies, the practice strategies, and the learner guidance strategies that are required and appropriate to promote acquisition of the instructional goal.
In ITT instructional strategies represent various ways to show, or request the student to provide, the elements of knowledge objects. Hence, an instructional strategy is an algorithm for processing the knowledge data (elements) of knowledge objects.
The buttons that appear below will be found at the top and bottom of each page of the discussion. The first button will take you back to the previous page (in this case, to the beginning of paper #22). The middle button will take you to the ITForum home page. The last button takes you forward into the remaining sections of the paper.
