ABSTRACT
The lower part of the figure is identical to the schematic training blueprint presented in Chapter 2 (Figure 2.1) and contains the four activities that cor respond with the four blueprint components. The design o f learning tasks is the heart of this training blueprint. For each task class, learning tasks are designed that provide learners with variable whole-task practice at a particular com plexity level until they reach the pre-specified standards for this level, where upon they continue to the next, more complex task class. The design o f supportive information pertains to all information that may help learners carry out the problem-solving, reasoning, and decision-making (i.e., non-recurrent)
aspects of the learning tasks within a particular task class. The design of proce dural information pertains to all information that exactly specifies how to carry out the routine (i.e., recurrent) aspects of the learning tasks. And, finally, the design of part task practice may be necessary for developing selected to-beautomated recurrent aspects to a very high level of automaticity. The two activities on the central axis of the figure sustain the design of learning tasks. At the top, the development of assessment instruments makes it possible to determine to which degree learners have reached pre-specified standards for acceptable performance. Because complex learning deals with highly integrated sets of learning objectives, the focus is on the decomposition of a complex skill into a hierarchy describing all aspects or constituent skills relevant to performing real-life tasks. Assessment instruments should make it possible to measure performance on each of these constituent skills, and to monitor learners’ progress over learning tasks, that is, over time. At the center of the figure, the *sequencing* of learning tasks describes a simple-tocomplex progression of categories of tasks that learners may work on. It organizes the tasks in such a way that learning is optimized. The simplest tasks are linked to the entry level of the learners (i.e., what they are already able to do when they enter the training program) and the final, most complex tasks are linked to the final attainment level of the whole training program. In adaptive learning based on frequent assessments, each learner receives a unique sequence of learning tasks adapted to his or her individual learning needs. In on-demand education, learners are able to select their own learning tasks but will often receive support and guidance in doing so (i.e., secondorder scaffolding). The two activities on the left side of the figure, analyzing cognitive strategies and analyzing mental models, sustain the design of supportive information. They are drawn next to each other because they have a bi-directional relationship: one is not conditional to another. The analysis of cognitive strat egies answers the question: How do proficient task performers systematically approach problems in the task domain? The analysis of mental models answers the question: How is the domain organized? The resulting *systematic approaches to problem solving* (SAPs) and *domain models* are used as a basis for the design of supportive information for a particular task class (see Chapter 7.1). There is a clear reciprocity between the sequencing of learning tasks in simple-to-complex task classes and the analysis of non-recurrent task aspects: More complex task classes require more detailed and/or more embellished cognitive strategies and mental models than simpler task classes. The two activities on the right side of the figure, analyzing cognitive rules and analyzing *prerequisite knowledge*, sustain the design of procedural information and part-task practice. They are drawn on top of each other because they have a conditional relationship: Cognitive rules require the availability of prerequisite information. The analysis of cognitive rules identifies the condition-action pairs that enable experts to perform routine aspects of
tasks without conscious effort (IF condition THEN action). The analysis of prerequisite knowledge identifies what experts need to know in order to correctly apply those condition-action pairs. Together, the results of these analyses provide the basis for the design of procedural information (see Chapter 10.1). In addition, identified cognitive rules are precisely those rules that need to be automated through part-task practice. As indicated by the arrows in Figure 3.1, some activities provide preliminary input for other activities. This suggests that the best order for performing the activities would, for example, be to start by developing the necessary assessment instruments, then to continue by analyzing the non-recurrent and recurrent aspects and sequencing the learning tasks, and to end with designing the four blueprint components. Indeed, the ten activities have previously been described in this analytical order (e.g., van Merriënboer & de Croock, 2002), but in real-life design projects each activity affects, and is affected by, all other activities. This makes it an open question as to which order for carrying out the ten activities is most fruitful.
