According to Kirschner et al. [30], worked examples constitute the epitome of strongly guided instruction and typically consist of a problem formulation, solution steps and the final answer itself. Worked examples make visible an expert’s problem-solving schemata to explain the steps of a solution for novices. Novices are provided with a worked example, as a schema of how to solve a problem, so they avoid having to engage in unnecessary trial and error processes [23, 24]. It also allows novices to circumvent most of the limitations of short- term memory, as their attention is focused and the relationship between problem solving steps is demonstrated.
Studying worked examples is an effective instructional strategy to teach complex problem-solving skills, it can be more effective than learning by problem solving [25]. Providing novices with worked examples rather than problem-solving exercises was shown to be a particularly powerful and efficient method for enhancing novices’ learning [25]. This can be explained by CLT. According to Kirschner et al. [30], solving a problem requires problem-solving search and search must occur in our limited working memory. Problem-solving search is an inefficient way of altering long-term memory, because its function is to find a
problem solution, not alter long-term memory. Problem-solving search can overburden limited working memory resources to be used for activities that are unrelated to learning (e.g. extraneous cognitive load). In contrast, studying worked examples both reduces working memory load because search is reduced or eliminated, and directs attention to learning the essential relations between problem-solving steps. Students learn to recognize which steps are required for particular problems, the basis for the acquisition of problem-solving schemata.
Reed and Bolstad [42] indicate that one example may be insufficient for helping students to induce a usable idea, and that the incorporation of a second
example, especially one that is more complex than the first, increases students’ learning outcome significantly. Therefore, the more worked examples students see; the more experience they have to build up their problem-solving skills.
Atkinson et al. [29] suggested providing multiple examples (at least two) of solving the same problem type improves learning and transfer before practising. Later on, examples and practice should be intermingled. This is further
supported by Trafton and Reiser [40] who indicate that each worked example offered should be followed by a similar practice problem, since pairing each worked example with a practice problem produces better outcomes than a block series of worked examples followed by a blocked series of practice problems.
Sweller and Cooper [39, 43] demonstrated that learning from worked examples can be more effective than learning by problem solving, based on the
experimental results that students learned more by studying algebra worked examples than by solving the equivalent problems. The finding can be explained by using the CLT. First, when a student, particularly one with low prior
knowledge, reviewed an example, it helped to lessen cognitive load and
maximize initial learning. Second, the cognitive schema created by the student while he or she was studying the example can then be used to deal with an isomorphic or similar problem to solve, for example, one with similar structure or elements to the example. The student can then easily recall the similar, just- reviewed example and does not need to grapple with many new and unfamiliar details in solving the new problem and searching through memory. At the same time, students can keep in mind active cognitive processing to strengthen the understanding of this type of problem in order to achieve deep learning.
Chi et al. [44] found that examples drawn from college level physics textbooks often do not include all of the reasons why a certain step in the solution was performed. So learners needed to work it out themselves and this phenomenon was termed the self-explanation effect. They discovered learners tried to establish a rationale for the solution steps by pausing to explain to themselves the examples and those learners appeared to learn more effectively than those who did not exhibit this behaviour. Self-explanations, formulating the unwritten steps of an example or concept, helped students understand examples and problems. Chi [45] suggested that the self-explanation effect was actually a dual process, one involving the generation of inferences and the other is repairing the learner’s own mental model. If there is a divergence between the learner’s own mental representation and the model conveyed by the text passage or example solution, the learner will update his own mental model.
Atkinson et al. [29] claimed that self-explanations were an important learning activity during the study of worked examples. Unfortunately, according to Renkl’s research results[46], the majority of learners’ self-explanation occurred in a passive or superficial way, they spent very little time studying the examples and missed the opportunities to self-explain. The minority of successful learners, however, seemed to use different self-explanation styles, for example,
principle-based explanations, example comparisons or anticipative reasoning. To assist this, the learners should be guided to actively self-explain worked
examples.
Chi [47] also found that sometimes students’ self-explanation can lead to misconceptions, where incorrect causations were constructed. To avoid this, completed worked examples, which contain the explanations of the reason for every step, are desired and required. This ensures novices gain the right
explanation through the worked examples and is the key point to help them to construct schemata.
According to Renkl et al.[48], direct training in self-explaining appeared to be effective, as are structural manipulations of examples such as adding sub goal labels, utilizing an integrated format( e.g. integrating text and diagram or aural and visual information), or using "incomplete" examples. Therefore, Roy and Chi [49] suggested self-explanation as a trainable learning strategy needs to be
considered when designing worked examples. Specifically, learners are encouraged to bring their prior knowledge to bear on the interpretation of materials and test their evolving understanding.
Therefore, worked examples should encourage learners to undertake self- explanation activities, in a way that avoids their skipping over the partially- complete sections. They should be designed to assist learners to think inwardly and reflectively. Self-explanations for learners, especially for novices, should be scaffolded at different levels of skill acquisition and schemata construction, in order to support a greater variety of learners. Faded worked examples
represent a formalisation of the self-explanation principle, and are described next.