Learning factories represent a complex simulation of actual work and production processes. Target groups for this real, dynamic production environment can include students, apprentices, professionals or individuals already in the workforce. Abele et al. (2019) and Faßhauer et al. (2021) identified four essential characteristics that must be fulfilled by a learning factory. These include the depiction of authentic work processes considering technical and organizational aspects, a dynamic work environment corresponding to a real value chain, the actual ability to produce a product in the learning factory and a comprehensive didactic concept for the integration of respective teaching and learning arrangements. Learning factories thus prepare learners for the participation and shaping of complex and dynamic work processes in a connected and digitized working world.

The focus is not only on developing primary competencies in handling specific tasks and working with objects but also on the training and acquisition of, among other things, teamwork competence in collaboration within the learning factory (also between different professions). Process competence is also addressed to understand and potentially shape the connections in the production process through the entire value creation process (Weyand et al., 2023; Petrusch et al., 2023) .

As comprehensive as the challenge is for all involved actors, the actual gap in the research on this topic is surprising. While a number of studies address the concept of learning factories and their integration into respective education (see, for example, Faßhauer et al., 2021; Scheid, 2018; Soori et al., 2023 ), there is still a lack of a fundamental, research-based approach to learning factories as complex learning environments and a new learning space for vocational education, considering as many characteristic factors as possible. This especially includes a deep perspective on digital work processes as learning spaces in the 21st century. At the same time, research results on criteria for the successful application of learning factories in their respective contexts (training, further education and university education) are also missing.

As part of a joint project in the Quality Offensive for Teacher Education, jointly financed by the state of Baden-Württemberg and the Federal Ministry of Education and Research (BMBF), the project “Integrated Technical and Economic Didactics (TWIND)” examines conceptual developments in the context of learning factories. Therefore, this contribution reflects on the concept of vocational learning factories and the associated learning spaces. In addition, the initial findings from a descriptive exploratory survey focusing on the operational process of the learning factory and the use/orientation of didactic components are presented.

The German “dual” training system has garnered a solid international reputation (Clarke et al., 2020; Oeben & Klumpp, 2021) . According to numerous experts, Germany’s success in exporting high-priced goods is attributed to its technically skilled workforce (Pilz & Wiemann, 2020) . A prominent characteristic of the dual training system is its twofold learning environment, encompassing vocational school and the company. This ensures the fusion of theoretical knowledge with hands-on experience, predominantly occurring at the company for three to four days per week. The training commences with a contractual agreement between the trainee and the company. Over the customary three-year training period, attending vocational school is obligatory (Cedefop, 2021) . Additionally, trainees receive a progressively increasing salary as they progress in their vocational training. More and more learning factories are established for training and teaching to maintain the high standards in the dual training system, as well as in developing the demanded skills. Zinn (2014: p. 23) described the learning factory as a concept in which learners have authentic opportunities to work on professional tasks with profession-specific tools in a realistic learning environment. Therefore the learning factory should make a business context imaginable, in which real working conditions are simulated for learners. It is not a simple theory-practice supplement, but a complex, demanding spatial and didactic-methodical conceptualization.

To meet this high standard, learning factories are modularly structured. This allows for flexible responses to changing training situations as well as to the altered requirements of training companies. Moreover, continuous development and didactic fine-tuning are possible.

The more didactically demanding and technically complex learning factories (smart factories) at vocational schools are increasingly playing a key role in translating the technical complexity and innovative character of Industry 4.0 into a vocational work situation under realistic conditions.

Learning sites in vocational education undergo a continuous process of change and adaptation (Harteis, 2022) . An expression of an innovative and flexible character is the integration of new learning sites, such as learning factories. Within the concept of learning factories, cyber-physical systems (CPS) can play a crucial role in innovating training to modern learning sites. The National Science Foundation (NSF, 2016) defined CPS as ‘engineered systems that are built from, and depend upon, the seamless integration of computational algorithms and physical components’ (NSF, 2016: p. 13) . The realm of technology has undergone substantial transformation in recent decades. Components that once existed solely in mechanical or electrical form, particularly those describing logic, control and decision-making, are increasingly taking on the characteristics of embedded systems and software, known as cyber elements. The acronym “CPS” is commonly employed to depict technical systems constructed through the seamless integration of computational algorithms and the physical components on which they rely (NSF, 2016) . In this context, “cyber” encompasses computers, software, data structures and networks that underpin decision-making within the system, while “physical” encompasses not only the components of physical systems (e.g., the mechanical and electrical elements of an automated vehicle) but also the physical environment in which the system operates, such as a production line or a modular learning factory. CPS is closely intertwined with terms such as the internet of things (IoT) and smart cities, as well as with the fields of robotics and systems engineering (Thiede et al., 2016) . Liu and colleagues (2019) listed intelligent industrial manufacturing, intelligent transportation and smart power grids as examples of CPS. The characteristic feature is the transformation of traditionally isolated, automated systems into modern, interconnected intelligent systems. The goal is to connect occupational actors, systems and physical assets to generate significant economic value as a synchronized network. CPS features complexity, dynamic variability and heterogeneity arising from interactions between cyber and physical subsystems (Lee & Sehia, 2016) .

Thiede et al. (2016) highlighted the opportunities provided by training in CPS. Such training serves as an alternative or complement to theoretical content in educational contexts and, with embedded learning factories, represents a promising didactic approach. As CPS gains importance in manufacturing, education and training are becoming key factors for successful implementation in companies. Using a simplified assessment system based on the components of a CPS, a tool was created for targeted use in learning environments on this topic. The approach aligns with corresponding processes and systems on an industrial scale, as well as with the required skills for development and operation (Gräßler et al., 2016) . The goal was to implement CPS elements as flexible components in the learning factory. The main focuses in this highly networked learning factory system (learning factories for research and qualification purposes) included the following elements (Gräßler et al., 2016) :

The concept included the communication structure, decentralized production control and the heuristics and algorithms of each device in the factory to enable an autonomous, decentralized production control system. Each device in the production system had the ability to collect the information necessary for its own operation and communicate its status to other devices.

The actual design of learning factories with CPS is a complex process requiring detailed approaches to various aspects. Studies on the design and construction of learning factories have highlighted the limitations of the static concept of learning factories (e.g. Abele et al., 2019; Kreß et al., 2021; McHauser et al., 2020 ). Gräßler et al. (2016) emphasized that no design approach focuses on the goal evaluation, scalability and temporal feasibility of the learning factory concept as a whole. To make optimal use of the potential of CPS in learning factories and to unfold the potential of learning factories as places for future dynamic workplaces, it is necessary to think about and plan learning factories in a modular and interconnected fashion.

The demand for learning factories and their associated didactic means is immense, reflecting the increasing complexity of work processes. The high technical complexity in production processes in the context of Industry 4.0 is accompanied by high demands on the technical understanding of learners and educators. To understand and comprehend the work and business processes in the context of Industry 4.0, skilled workers in the industrial-technical fields, and increasingly also in the commercial fields, are confronted with this process understanding (Faßhauer et al., 2021) .

In terms of design, construction, and potential development, Frielinck (2023) , Gräßler et al. (2016) and Kreß et al. (2021) identified a range of design approaches for classifying learning factories in the context of research and qualification. In this context, deficiencies in the majority of existing learning factories were identified, covering the process through implementation and the operation of the completed learning factory.

Sudhoff (2021) and Löhr-Zeidler et al. (2016) identified over 20 fundamental focal points for learning factories in vocational schools that can be represented in a (complete) facility. The range extends from digitization processes to production management processes, automation, artificial intelligence, product development, production planning and plant control through a manufacturing execution system (MES). However, the topics are not always linked to professional actions; questions about understanding and categorizing technology often dominate.

In addition, there are questions about the basics of product development, work planning, production control, production management and software engineering for production systems, which can be implemented within the learning factory when creating a product. The connection between product development and subsequent production is understandable (Gräßler et al., 2016) .

Thus, learning factories combine a series of learning-promoting factors (Frielinck, 2023) that significantly influence knowledge transfer (Tynjälä et al., 2021) . A didactically valuable learning environment based on the processes of connectivity between education and work (Zitter et al., 2021) includes, according to Rau (2004) , the following skills:

The requirements for the development of competence in work have been determined to be predictability, task variety and feedback on results (Rau, 2004) . The demands placed on vocational learning factories, as well as the associated challenges for knowledge transfer, reveal the complexity of this relatively new learning space. The survey described below aimed to lay the foundation for a research-based approach to vocational learning factories as complex learning environments and new learning spaces for vocational education, taking into account as many characteristic factors as possible. This makes learning factories a valuable environment, providing a deepened perspective on digital work and business processes as learning spaces in the 21st century (Mueller et al., 2023) .

Learning factories are considered promising and methodically complex learning spaces (Öztürk et al., 2022) . However, there are currently few studies that investigate the use, orientation and success conditions of learning factories in vocational education.

The descriptive approach aimed to provide initial insight into learning factories as learning spaces. The descriptive analysis provided answers to questions such as which didactic components were extensively used, how the operation of the learning factory and complex teaching-learning arrangements interacted, what the actual spatial and technical equipment of the learning factory was, what possibilities the learning factory had to meet the requirements of Industry 4.0 and what level of utilization and acceptance a learning factory experienced within the vocational school. Additionally, aspects of school personnel development, especially the recruitment and training of teachers for the use of learning factories, played a role.

This paper reflects on learning spaces and the development of learning factories in vocational education, with the aim of establishing an initial scientific foundation for an exploratory descriptive survey on learning factories. Learning factories are considered promising and methodically complex learning spaces (Leppert, 2021) , yet their components and applications have barely been empirically investigated on a broad scale (Windelband et al., 2022) . The presented instrument represents a first step towards the gradual description and examination of learning factories in vocational schools. The main part of the questionnaire focused on the design of learning factories, didactic-methodical implementation in the learning process, the process design of the learning factory and the degree of digitalization.

Aspects of school personnel development, especially the recruitment and training of teachers for the use of learning factories, were also examined. A gauge for the achieved level of innovation was cross-departmental in-school collaboration, as well as learning environments with dual partners in the context of the learning factory. This was achieved in the questionnaire through items related to the conditions of the learning factory and additional questions about the dual partners involved in the design and implementation of the learning factories. This also included descriptive parameters such as the company size of the involved partners. In addition, there were several basic questions describing the equipment and conditions of the learning factory. These included self-assessment items (from 1 = completely inadequate to 5 = completely sufficient) for the following fields:

The presented initial results show how dynamic and heterogeneous the process of implementing learning factories is. Many learning factories are still in their initial phase. New paths are being considered, established ones are being proven and they are being compared with the requirements of the constantly digitizing working world. The surveyed actors in vocational schools quickly encountered formal and substantive limits. The results clearly show that the vocational fields in the context of an Industry 4.0 world are not always reflected. One study participant vividly summarized this: “The use is overall difficult; there is a lack of ideas/concepts. The system is expensive, extensive, and yet has little to do with practice.” This statement vividly illustrates that the implementation of a learning factory can represent a significant innovation boost for individual vocational schools (Windelband et al., 2022) . However, this success is only achieved if all involved parties share the same goal and are equipped with sufficient resources and competencies. The results suggest that significant conflicts are expected, especially in the conception phase and the initial phase of operating the learning factory. The introduction of a learning factory is a litmus test that questions existing structures and sometimes forces new structures. The necessary factors for successful knowledge transfer highlighted by Rau (2004) and Tynjälä et al. (2021) , such as involvement of one’s own ideas in task execution, predictability, and experience of control over a situation, also apply to the actors of the learning factory. The requirements posed to and the challenges associated with knowledge transfer in learning factories reveal the complexity of the learning space. This includes the relationship between learning in basic laboratories and learning on the entire system and, thus, actually in the learning factory. The research hypothesis here was that training primarily takes place in the basic laboratories in vocational schools, and it is only successful for a few apprenticeships (e.g. mechatronics) and in technician classes to implement adequately problem-oriented vocational learning situations throughout the entire learning factory. The results confirm the findings of earlier exploratory studies on the use of learning factories in vocational schools in Baden-Württemberg, which were conducted at a very early stage of implementation (e.g. Roll & Ifenthaler, 2020; Scheid, 2018; Windelband, 2023; Zinn, 2014 ).

An important factor for the success and sustainable integration of the learning factory into everyday vocational life is to integrate it firmly into teacher training and further education. If this succeeds, learning factories can make a significant contribution to vocational education in the technical and commercial professions and in other vocational disciplines. They promote thinking, working, and cooperating in interconnected production systems and have an overview of the entire value-added process.

The nature of this descriptive exploratory study can be seen as a limitation (Cohen et al., 2018) . The lack of fundamental and comprehensive research for the comparison and validation of the initially presented results complicates their classification. Nevertheless, the study represents a first step in categorizing previously neglected areas in vocational learning factories into a larger context. This can only be considered the first point of an emerging project-based research series. The planned follow-up studies will develop and implement an actor-specific questionnaire based on the descriptive results.

The concept of self-assessment by surveyed individuals has been extensively discussed in the research literature (see Cohen et al., 2018; Flick, 2022 ). Limitations of self-assessment usually revolve around the argument of a lack of control over answers by, for example, superiors or experts in pedagogical-psychological diagnostics. In this context, a series of points used in the survey were mentioned as a solution. These include conducting surveys entirely anonymously, ensuring that participating individuals incur no disadvantage through participation or nonparticipation, considering participating individuals as experts in the treated field and formulating questions that do not allow conclusions about socially desirable answers. Cohen et al. (2018) emphasized that in this context, no method of self-assessment can guarantee absolute truths. However, the meaningfulness of self-assessment questionnaires can be significantly increased with the conscious handling of measurement-theoretical sources of error.

The implementation of a learning factory represents a significant innovation boost for individual vocational schools, which goes well beyond actual technological advancement. It is simultaneously a result and an occasion for a potentially profound social design process within the framework of school development, influencing at least organizational development, personnel, and teaching development, as well as framing internal and external cooperation. However, a significant influencing factor is the technology itself, including the composition of the technical components of the learning factory, the level of digitization of the school and the overall use of digital media (Windelband et al., 2022) .

The results of the research on learning factories and the initial descriptive results will be incorporated into the design of a second survey. Validated scales on teaching-learning contexts in vocational education, knowledge transfer and learning factories as innovators in vocational education will be combined with the described set of items for the descriptive classification of processes at a learning factory. At the same time, the circle of vocational schools to be surveyed will be expanded to additional federal states to obtain a comprehensive assessment of the development of vocational schools in the context of the requirements of Industry 4.0 and digitization.

The article processing charge was funded by the Baden-Württemberg Ministry of Science, Research and Culture and the University of Education Schwäbisch Gmünd in the funding programme Open Access Publishing.