Building a STEM Training Program for Preschool Educators: Feasibility Study

A new study published in the International Journal of Early Childhood offers compelling evidence that well-designed professional development can transform how early childhood educators teach science, technology, engineering, and mathematics. Researchers at the SEAMEO Regional Centre for QITEP in Science in Bandung, Indonesia, developed and rigorously evaluated a STEM training program for teachers of young children, combining two established frameworks: the ADDIE instructional design model and the CIPP evaluation model. Their findings, published as volume 58 of the journal went to press in April 2026, show measurable gains in teacher knowledge, strong alignment between training materials and classroom needs, and, most strikingly, demonstrable changes in how young children engage with engineering challenges.

The research team, led by Lintang Ratri Prastika alongside Reza Setiawan, Zuhe Safitra, and Apriyagung, set out to address a persistent gap in early childhood education. While STEM learning has become a global priority from primary school upward, the earliest years of schooling have often been overlooked, in part because many early childhood educators lack confidence in science and mathematics content. Previous studies have documented that preschool teachers frequently report low self-efficacy in science instruction, and that professional development opportunities tailored to the specific demands of teaching four- to six-year-olds remain scarce. The Indonesian context made this problem particularly urgent: the country’s national curriculum for early childhood, the Foundation Phase, now explicitly calls for science process skills and problem-solving competencies, yet few training programs have been designed and validated specifically for the educators responsible for delivering them.

The methodological architecture of the study is one of its most distinctive features. Rather than treating program development and program evaluation as separate enterprises, the researchers fused the ADDIE model, which stands for Analyze, Design, Develop, Implement, and Evaluate, with the CIPP framework, an acronym for Context, Input, Process, and Product evaluation pioneered by educational evaluator Daniel Stufflebeam. In practice, this meant that each stage of ADDIE was paired with a corresponding evaluative lens. During the analysis phase, the team conducted a context evaluation, surveying teachers about their existing knowledge, challenges, and needs. The design and development phases were informed by input evaluation, examining whether the chosen materials, experts, and delivery formats were appropriate. The implementation phase was monitored through process evaluation, capturing how teachers engaged with the training in real time. Finally, the product evaluation assessed outcomes for both teachers and their students.

The training program itself followed an “In-On” model, a structure that alternates between in-service workshops and on-the-job application. Teachers first attended intensive sessions where they learned the theoretical foundations of STEM education for young children, including how to frame age-appropriate engineering challenges and how to guide children through the Engineering Design Process, commonly abbreviated as EDP. The EDP is a cyclical problem-solving sequence widely used in engineering education, typically encompassing stages such as asking a question, imagining solutions, planning, creating a prototype, testing it, and redesigning based on what the test reveals. After the workshops, teachers returned to their classrooms to implement STEM activities with their students, with the research team observing and collecting data on both teacher practice and child engagement.

The results on the teacher side were encouraging. The program was categorized as both feasible and effective, with normalized gain scores, a standard metric in science education research known as N-gain, of 0.41 and 0.38 across the measured cohorts. In the classification system popularized by physics education researcher Richard Hake, N-gain values in this range fall into the “medium” category, indicating meaningful improvement in teacher knowledge that goes well beyond what would be expected from chance. Perhaps equally important, the relevance of the training materials to teachers’ actual needs was rated between 94 and 100 percent, suggesting that the context analysis conducted at the outset of the project successfully identified the specific gaps the training needed to fill. This is a critical finding for program designers, because professional development initiatives frequently fail not because the content is wrong but because it does not match what practitioners actually face in their classrooms.

The most consequential findings, however, concern what happened to the children. When trained teachers implemented STEM activities in their classrooms, the research team documented the degree to which children engaged with each stage of the Engineering Design Process. Participation was highest in the Create stage, at 97 percent, followed closely by the Test stage, at 90 percent. These figures indicate that once children had planned their solutions, they were almost universally willing and able to build prototypes and subject them to real-world trials, whether that meant constructing a tower from recycled materials, designing a water channel, or building a simple machine from everyday objects.

But the single most interesting number in the study may be the 71 percent implementation rate for the Redesign stage. In engineering education, redesign is often considered the heart of the discipline: the willingness to see a failed or imperfect prototype not as an endpoint but as information, and to iterate accordingly. For young children, this stage carries particular developmental weight, because it cultivates resilience, the capacity to tolerate frustration and persist through setbacks. The authors highlight this finding as evidence that the program was highly effective in building student resilience through the Redesign stage. A 71 percent rate means that roughly seven in ten children, having watched their creation fail a test, chose to go back and try again rather than abandon the task, a behavioral pattern that educators and psychologists widely regard as foundational for later academic persistence.

The study’s implications extend beyond Indonesia. Globally, early childhood STEM education has been expanding, with systematic reviews documenting a growing body of empirical work on STEM in the years before formal schooling. Yet the field has struggled with a chicken-and-egg problem: children benefit from early exposure to engineering thinking, but their teachers often lack training in how to facilitate it, and teacher training programs in turn lack evidence about which designs actually change classroom practice. By integrating a rigorous evaluation framework directly into the development process, the ADDIE-CIPP approach demonstrated in this study offers a template for how training programs can be built, tested, and refined in a single coherent cycle, rather than developed first and evaluated, if at all, years later.

The authors are candid about the limitations and the work that remains. Their strategic recommendations focus on two priorities. First, the language of the training modules should be simplified, an acknowledgment that dense academic phrasing can create barriers for practitioners who learn best from concrete, classroom-ready guidance. Second, and more ambitiously, they call for in-depth analysis of how teachers’ skills actually transform as they direct science processes, master effective questioning techniques, and facilitate the complete EDP cycle. Effective questioning is a subtle craft in early childhood settings: the difference between asking a child “Why did your tower fall?” and “What could you change so it stands taller?” can determine whether a child engages in genuine scientific reasoning or simply seeks adult approval. Understanding how teachers acquire and deploy these skills over time, the authors argue, is essential for supporting sustainable teacher professional development rather than one-off workshops that fade within weeks.

The study also situates itself within a broader shift in how scientists and educators think about early learning. Developmental research over the past two decades has steadily dismantled the assumption that young children cannot engage with engineering or systematic problem-solving. Studies of play-based learning have shown that preschoolers naturally experiment, hypothesize, and iterate when given appropriate materials and scaffolding. What has been missing is a workforce of educators equipped to recognize and extend these spontaneous scientific behaviors. Programs like the one developed in Bandung suggest that the bottleneck is not children’s capacity but adult preparation, and that targeted, needs-responsive training can move the needle substantially within a single professional development cycle.

For education policymakers, the study’s feasibility findings carry a practical message. A program in which materials matched teacher needs at rates approaching 100 percent, and which produced medium-category knowledge gains, was achieved within a structured development cycle that other regions could replicate. The explicit pairing of ADDIE with CIPP provides an audit trail: funders and ministries can see not only whether a program worked but why, at which stage, and with which populations. In an era when education budgets are scrutinized more tightly than ever, that kind of mechanistic accountability may prove as influential as the headline outcomes.

The research, published on 22 April 2026 in the International Journal of Early Childhood, was conducted by a team affiliated with SEAMEO, the Southeast Asian Ministers of Education Organization, whose regional science education center in Bandung specializes in improving science teaching quality across Southeast Asia. The authors declare no conflicts of interest. As early childhood STEM education continues its expansion from niche interest to mainstream policy priority, this study provides both a proof of concept and a methodological blueprint: design training around teachers’ real needs, embed evaluation at every stage, and measure success not only by what teachers know but by what children actually do in the classroom, tower by tower, test by test, and redesign by redesign.