(van Merriënboer 2020) described Learning tasks as whole tasks that should promote learning processes. Above all, learning tasks should prepare students for their future work. The aim is learning through doing and concrete experiences. Learning tasks should always be complete tasks that integrate knowledge, skills, and attitudes, which is why these tasks are also referred to as “whole tasks” (van Merriënboer, 2020). To achieve this, learning tasks should be variable, have various levels of complexity, and learners should be guided to solve them. At the beginning of a learning process students should receive support and guidance for difficult and complex tasks and should be led to the next level of complexity if they can solve a task without any help. Support is gradually reduced until students can solve the task themselves. This process of creating a fully support and reducing the support gradually is called Scaffolding (van Merrienboer et al., 2003). Support can take different forms, such as examples of solutions, completing incomplete solutions and finally solving the complete task (Renkl and Atkinson, 2003)—and can also be provided by the instructors themselves or externally through worksheets with guiding questions, etc. For this, learning tasks consist of both routine and non-routine aspects. Routine tasks are always done in the same way. Non-routine tasks require problem solving, weighing up and making decisions (van Merriënboer, 2013). Learning tasks can encompass anything students engage with—such as cases, projects, work assignments, problems, or other activities. Within each complexity level, the tasks should be varied in a similar way as they might vary in real-life tasks, and typically, learning journeys should begin with simple tasks and gradually progresses to more complex ones over time (van Merriënboer and Sweller, 2010).

Learners need information to learn profitably from the learning tasks (van Merrienboer and Kirschner, 2012). According to (van Merriënboer 2020) supportive information help learners to deal with non-routine aspects of a learning task such as problem solving or decision making. Ideally supportive information connects the existing prior knowledge of the learners with the knowledge the learners should know in order to be able to perform such non-routine tasks. Therefore, information should be structured in a way that prior knowledge can easily be linked to the new information. Cognitive feedback is very important here (Hattie and Timperley, 2007; Kluger and DeNisi, 1996). Through feedback, learners are able to compare their cognitive schemata with others and correct them if necessary. Supportive information is often represented as theory from books, lectures, or online resources and should be identical for all learning tasks of a certain level of complexity. Such supporting information can be presented at the beginning of a new unit as well as during working process on the unit. For the next level of complexity, supportive information needs to be extended and supplemented (van Merriënboer, 2020).

In addition to the non-routine aspects, there are also the routine aspects connected to real-life tasks. For that, according to (van Merriënboer 2020), procedural information should assist learners with such routine aspects. The aim of procedural information is to enable students to develop cognitive rules that become cognitive schemata through practice so that they can be used quickly, easily and error-free. As procedural information is therefore usually provided during the work on a learning task, they are also called just-in-time information. Typically, information is presented as a “How-to” or “Step-by-step” instruction—given either by the teacher themselves or in the form of a user guide. Learners should get procedural information at the beginning of a new task, while with the next comparable tasks the procedural information is gradually reduced. It is important to formulate information at a basic level so that every student, including lower-achieving ones, are able to understand them (van Merriënboer, 2020).

Learning tasks consists of both routine and non-routine aspects. Typically, routine aspects are practiced within the context of learning tasks. However, if solving a learning task requires a high degree of automation of certain routine components—and these components are not sufficiently trained while solving the learning task itself, students should engage in part-task practice. According to (van Merriënboer 2020) the purpose is to develop fully automated procedural schemata (Anderson, 1982), which is why in cases where working on the complete learning task does not provide enough repetition for automation, specific routine skills should be practiced separately in part-tasks. Ideally part-task practice is integrated into the work on whole learning tasks process and aligned with the complete task. To be effective, part-tasks should begin within an appropriate cognitive context. For instance, the routine elements should be made explicit within the full task before isolating them for practice.

In addition to the four components that should enable competence-oriented learning, the model also lists 10 steps of design activities (van Merrienboer and Kirschner, 2012) to design courses fitting to the 4C/ID model. Within the 10 steps, the four components (steps 1, 4, 7 and 10) form a kind of super-category. The steps in between concretize the respective superordinate category. The first component “Learning tasks,” for example, is further concretized with the help of steps two and three. We consider a detailed description of the 10 steps in a published article as an indicator of successful implementation of 4C/ID—and summarize which information regarding each of those steps could be expected in Table 1.

In a globalized world, circumstances are constantly changing. Professional pre-service training programs therefore cannot prepare people for a career that lasts for decades but must be complemented by regular further training (International Labour Office, 2018). For example, teaching and school practice are in a constant state of flux, with new challenges emerging all the time. A study program cannot fully prepare students for the changing conditions of the next 40 years (Lipowsky and Rzejak, 2017), as training to become a good teacher is a complex task that ideally combines theory and practice. 4C/ID teaches skills in real-life situations and therefore seems ideal for further training in professional contexts—with each of the four components addressing specific needs for in-service training.

Professional development as in-service training is essential when professional actions are complex which 4C/ID a challenge that the 4C/ID model addresses by designing learning tasks that mirror authentic challenging situations. Tasks should be variable so that the tasks remain manageable. Complexity increases steadily over the task progress. Holistic tasks such as learning tasks, develop and use knowledge, skills and attitudes in equal measure. In addition, tasks are performed realistically and holistically, which prevents a separation of partial competences. This is an important skill that is highly relevant in a wide range of occupations.

The 4C/ID model is widely regarded as a valuable framework for curriculum design in vocational education and training. One essential difference of pre-service and in-service training is the learners prior experience with the specific competencies in practice. This knowledge facet cannot be ignored as it can be particularly useful or lead to overgeneralizations of personal experience—which is why in-service training should especially take learners prior experiences into account. Supportive information in the 4C/ID model helps to link the new-to-learn knowledge to existing prior knowledge—and prior experiences—and make it available in connection. It also prevents the learning of individual aspects in isolation. The link between knowledge and benefits facilitates the transfer of knowledge to practical situations.

Activities can be automated using procedural information and part-task practice. The rule-based automation of part-tasks can reduce the workload when completing the task. For learners with practical experience in particular, the aim is to reduce the workload of everyday tasks and to make it easier to work on complex tasks. This is why the application of the 4C/ID model can be particularly beneficial in training programs.

In summary, 4C/ID provides a structure for well-developed in-service professional development programs that is perfectly able to take into account the experiences of participants acquired during their execution of their jobs. 4C/ID teaches skills in authentic real-life situations and is therefore ideal for continuing professional development in professional contexts. Nevertheless, the question is whether the 4C/ID approach will also achieve the desired learning effects.

(Lipowsky and Rzejak 2021) describe, as the result of a research synthesis, several principles for the methodological and didactic design of effective (in-service) training programs—with teacher training in mind. The similarities between the Principles of effective teaching and the 4C/ID model are presented below.

As a first principle, (Lipowsky and Rzejak 2021) name the “linking of input, testing and reflection phases” (“Verknüpfung von Input-, Erprobungs- und Reflexionsphasen” S. 13). Knowledge acquisition, the actual action and reflection on the action carried out should not be realized separately, but as far as possible simultaneously in a joint task. A similar goal can be found in the 4C/ID model. Learning tasks should be as holistic as possible, in which knowledge, skills and attitudes are promoted in equal measure. These holistic tasks should be implemented in authentic everyday situations as far as possible, just as required by the principles for effective teacher training.

The next principle that emerged from the research synthesis was “the principle of feedback and coaching” (“Feedback und Coaching” S. 13). The aim is to support teachers' experiences and learning processes through feedback, examples and suggestions. In the 4C/ID model, this principle is implemented in the form of procedural information. Procedural information is often implemented in the form of just-in-time information—where teachers are provided with the information or examples they need during their work. Procedural information can also be implemented in the form of corrective feedback. In this case, the necessary instructions for the correct completion of the task should also be given during completion (Lipowsky and Rzejak, 2021).

As a final principle, (Lipowsky and Rzejak 2021) name “meaningful content and activities” (“Bedeutsame Inhalte und Aktivitäten” S. 13). The benefits and relevance of the training content should be emphasized through practical relevance. This principle is implemented in the 4C/ID model primarily through supportive information. Supportive information provides the theoretical background to the learning tasks. By working on holistic, authentic tasks, the benefits of theory become omnipresent. Supportive information can be accessed at any time while working on the learning task. This creates a link between theoretical content and practice, in line with (Lipowsky and Rzejak 2021).

By following the PRISMA guidelines (Page et al., 2021) we decided to use “4C/ID” OR “Four-Component Instructional Design” for our search syntax to search the Eric and Web of Science databases. In both databases, we searched in title and abstract. We assumed that the most relevant studies are those that explicitly refer to 4C/ID in the title or abstract, rather than just mentioning it somewhere in the article. In such studies, it is plausible to assume that essential parts of the theoretical background can be traced back to 4C/ID. This resulted in 38 hits in ERIC and 77 hits in Web of Science. In total, we found 115 texts. For these 115 texts we had strict inclusion and exclusion criteria. We excluded:

Of the 115 texts, 17 papers had to be excluded because they appeared twice or more. Six papers were excluded because access to the texts was not possible. Forty-four papers were excluded because it was not a training program, 2 papers were excluded because 4C/ID was not the theoretical basis. We also excluded 2 papers because of various reasons: one paper because it was too short to get differentiated information and one paper because the training was held on 4C/ID but was not methodically applied.

This left us with 38 texts that were ultimately included because all inclusion criteria are fulfilled (Figure 1). These inclusion criteria were:

The decision on whether studies were included was double coded, with coding performed by the first author of the paper and an additional trained researcher. The agreement between the two coders reached a fair agreement (Cohen's Kappa κ = 0.36)—but the relatively low value is why all studies for which at least one coder recommended inclusion were reviewed in a subsequent cycle and evaluated conclusively by the whole team. The relatively low level of agreement in the coding can probably be explained by the broad definition of professional development, which was the main criterion for the inclusion or exclusion of papers.

In a second step we performed a snowball backwards search. For each of these texts, we used the “Cited by” function in Google Scholar to identify texts that cited the 38 texts above We assumed that the 38 texts found could be prototypical regarding the use of 4C/ID. We therefore expected that other (possibly less prototypical) 4C/ID training studies would refer to the methodological design or other aspects of these 38 papers. It is therefore to be expected that these 38 studies will be cited by prototypically similar papers, among others. This is why we looked at the papers that cite them. These 38 texts were cited 971 times in total (on 21 November 2023).

Again, we excluded 202 texts because they appeared twice or more and 144 texts, we could not get access to (Figure 1). We also excluded 98 texts because they were not written in English or German. This left us with 528 texts where we have skimmed the abstract and texts. For these texts we had the same exclusion and inclusion criteria as for our first search. We excluded 350 texts because they do not describe a further training or the 4C/ID model. In addition, we excluded 156 texts because these were not articles but book contributions, books, dissertations, master's theses, PowerPoint presentations or Wikipedia entries. We also excluded 5 texts of other reasons (Figure 1). In the end we were left with 16 additional texts which we ultimately included. After all this we had in total 55 texts left which were coded in the following step.

In addition, all papers were analyzed qualitatively with regard to the implementation of the four components, the profession, the subjects of matter and the real-life tasks. Therefore, we focused our qualitative description on the specific implementation of the four components.

The analysis of the papers shows that the texts we consider relevant were published in the last 20 years, between 2004 and 2024. Since 2018, the average number of publications has increased eightfold. Figure 2 illustrates this trend.

Ideally, papers referencing the 4C/ID model as the foundation for their training should provide a more detailed account of its implementation. For clarity and comprehensibility, we recommend describing at least the model's four core components. However, a step-by-step outline of the training process, following the 10 steps of 4C/ID, would offer even greater insight and transparency. This is currently not the case for the articles reviewed in the present synthesis:

Of the total of 55 texts, 30 texts (54%) were categorized as “yes” and thus described the four components with reference to the training, whereas 26 texts (46%) did not describe the four components or only described them in general terms. Because the quality of the implementation of 4C/ID can only be derived from at least one description of the four components, or even better from the description of the 10 steps, it is difficult for the other 46% of the papers to assess the extent to which 4C/ID has been implemented at all.

For the 10 steps, we also analyzed whether the 10 steps were described with reference to the training program. A general description of the use of the 10 steps was not sufficient here. Six of the 56 (11%) texts were rated “yes” and therefore described the 10 steps in detail, while 50 texts described the 10 steps only in general terms or not at all.

Eleven of the 55 texts describe training programs for teachers. Seven papers describe at least the four components, two additionally describe the 10 steps in detail. Thematically, teachers in the training courses should learn how to design differentiated lessons (mentioned above) (Frerejean et al., 2021), integrate technology into learning and teaching (Lim et al., 2015), create research-based learning programs (Yan et al., 2012), create grade books with Excel (Lim et al., 2009), use visual media (Xu et al., 2020), implement assessment for learning in lessons (Wolterinck et al., 2022), improve foreign language skills (Larmuseau et al., 2018b; Wittmann and Olivier, 2021), provide differentiated instruction in lessons (Meutstege et al., 2023), conduct and publish research (Kukharuk et al., 2023) and conduct integrated lessons between languages and science (Rhodes et al., 2024).

The four components are implemented in a variety of ways in the training programs. Learning tasks range from concrete implementation at school (Frerejean et al., 2021) or in real-life situations (Lim et al., 2009) to implementation on online platforms such as Moodle (Lim et al., 2015). General information such as graphic representations of the 4C/ID principles (Frerejean et al., 2021), teaching/learning strategies (Yan et al., 2012) or structures to support lesson preparation (Rhodes et al., 2024) are provided as supportive information in the various training courses. Specific information such as concrete examples, checklists and handouts (Meutstege et al., 2023) are also provided. For procedural information, enquiries can be made to coaches (Frerejean et al., 2021) or step-by-step instructions are provided, e.g., for various Excel functions (Lim et al., 2009). None of the texts provided specific tasks for part-task practice.

In general, it was surprising to find only a few texts on in-service training with 4C/ID for teachers, especially because 4C/ID is just as suitable for teacher training as it is for medical or technical training and fulfills the criteria for effective training (Lipowsky and Rzejak, 2015). The real-life tasks of the individual training programs relate primarily to the technical and methodological skills of teachers. There are still gaps, particularly in subject-specific and didactic topics, although the implementation of these areas can also be interesting.

We give a short summary of the raised research questions and the corresponding answers based on our systematic review.

How is 4C/ID documented in publications that cite 4C/ID as the theoretical framework for development of in-service training courses?

In which professional groups is 4C/ID used and for which real-life tasks?

What is the estimated mean pre-post effect for such in-service 4C/ID interventions?

If 4C/ID is implemented, what methodological approaches and which methodological concretizations are used for components and steps?

Has 4C/ID been implemented in teacher training programs to date? And if so, how and for what?