Evaluating a Teacher Development Course for Teaching STEM

Activities with Introductory Internet of Things Concepts and AI Data

Model Training Skills Using the TPACK Framework:

Problem-Solving and Digital Creativity

Siu Cheung Kong

1,2 a

, Cora Ka Yuk Siu

and Wing Kei Yeung

Artificial Intelligence and Digital Competency Education Centre, The Education University of Hong Kong, HKSAR, China

Department of Mathematics and Information Technology, The Education University of Hong Kong, HKSAR, China

Keywords: AI Data Model Training, Digital Creativity, Internet of Things, STEM, Problem-Solving,

Teacher Development, TPACK Framework.

Abstract: We designed a 6-hour teacher development course aimed at enhancing teachers’ competency in teaching

STEM activities. The course focused on teaching teachers how to develop learners’ problem-solving abilities

and digital creativity using both introductory concepts of the Internet of Things (IoT) and artificial intelligence

(AI) data model training skills in teaching STEM activities. This study evaluated the teachers’ competency in

teaching STEM activities and the outcomes of their creative ideas in solving problems using what they had

learned in this course. Two hundred and one teachers from 108 primary schools attended the course, of whom

191 responded to the pre- and post-course surveys on the TPACK framework, and 176 of them produced

artefacts demonstrating their digital creativity. The paired t-test results indicated statistically significant

improvement on all 17 TPACK items, with a large effect size (Cohen’s d = 1.213). In the digital creativity

evaluation, 82.20% of the teachers demonstrated digital creativity and expressed their ideas in designing

introductory IoT systems, and 72.77% of the teachers included AI components in their design. One future

research direction is to evaluate primary students’ learning outcomes in STEM activities with these

introductory concepts of IoT and AI data model training skills.

1 INTRODUCTION

In the dynamic context of today’s rapidly advancing

technological landscape, smart systems are assuming

an increasingly prominent role. Such systems

seamlessly integrate artificial intelligence (AI) and

Internet of Things (IoT) with a diverse array of

sensors and actuators. By collaboratively collecting,

analysing, and responding to data, these components

enable smart systems to operate autonomously (Al-

Fuqaha et al., 2015).

The integration of AI and IoT into STEM

education can significantly enhance students’

problem-solving skills and ignite their digital

creativity, empowering them to solve real-life

problems (Kong et al., 2024). It is thus imperative that

https://orcid.org/0000-0002-8691-3016

https://orcid.org/0000-0001-8176-2880

https://orcid.org/0009-0009-3845-7448

primary students are provided with more

opportunities to understand and engage with today’s

AI-permeated world (Kim et al., 2021).

Although research has addressed the development

of students’ AI literacy (Touretzky et al., 2019), the

integration of AI and IoT within STEM education

remains underdeveloped (Kong et al., 2024). As an

emerging technology, AI can induce anxiety, which

may hinder its future application and behavioural

intention in professional contexts (Wang & Wang,

2022). Educators play a key role in shaping their

students’ futures by imparting imperative knowledge.

To effectively teach and disseminate this new

knowledge, they must possess a profound

understanding of the subject matter (Hsu et al., 2023).

Many educators currently exhibit low self-efficacy in

Kong, S. C., Siu, C. K. Y. and Yeung, W. K.

Evaluating a Teacher Development Course for Teaching STEM Activities with Introductory Internet of Things Concepts and AI Data Model Training Skills Using the TPACK Framework:

Problem-Solving and Digital Creativity.

DOI: 10.5220/0013201100003932

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Conference on Computer Supported Education (CSEDU 2025) - Volume 1, pages 36-47

ISBN: 978-989-758-746-7; ISSN: 2184-5026

participating in in-service teacher education

programmes and feel unprepared for AI education.

Therefore, enhancing their understanding and

mastery of AI applications is essential (Hsu et al.,

2023). Additionally, it is important for teachers to

comprehend AI concepts, solve problems using AI,

be psychologically ready to utilise AI, and understand

the ethical dimensions of AI problem-solving (Kong

& Yang, 2024).

We therefore propose an innovative approach that

incorporates these concepts into primary school

STEM activities. Our design focuses on enhancing

students’ understanding by facilitating the creation of

physical artefacts using introductory IoT concepts

and AI data model training skills, thus bridging the

gap between theory and practical application.

The IoT and AI are increasingly shaping our

world and are poised to become an integral part of

everyday life in the digital era (Al-Fuqaha et al.,

2015). Therefore, it is crucial to introduce these

concepts to primary students. We have selected AI

and the IoT as the foundation for our STEM activities.

By thoughtfully embedding introductory experiences

in AI data model training into educational content, we

can include AI data model training and interaction in

STEM activities. This integration is vital for helping

students understand how automated systems are

designed, thereby enriching these systems’

complexity and enhancing their usability and

accessibility. Moreover, these activities can foster

digital creativity by enabling students to consider

applying these concepts in new contexts.

However, in-service teachers often have limited

technological background related to STEM and lack

confidence in teaching these subjects; therefore,

teacher development is crucial to successfully

integrating STEM education into classrooms

(Cavlazoglu & Stuessy, 2017; Lo, 2021). This study

therefore seeks to answer the following questions:

(1) Which components of the TPACK framework

(i.e., content knowledge [CK], technological

knowledge [TK], pedagogical knowledge [PK],

pedagogical content knowledge [PCK],

technological content knowledge [TCK],

technological pedagogical knowledge [TPK],

and technological pedagogical content

knowledge [TPACK]) show the greatest

improvements after teachers complete a

development course?

(2) How do teachers benefit in terms of their

creativity after attending this course, as

evidenced by their design artefacts?

(3) What are the most valuable components of the

teacher development course?

2 LITERATURE REVIEW

2.1 STEM Education: Fostering

Students’ Problem-Solving Ability

and Digital Creativity

STEM education plays a key role in cultivating

students’ problem-solving skills. Research indicates

that students are more engaged in STEM activities

when they create artefacts using technology (Hanif et

al., 2019). STEM education encompasses various

aspects, such as causal reasoning, sequencing,

conditional reasoning, and engineering systems

thinking (Sullivan & Heffernan, 2016). These skills

are essential for understanding and addressing

complex problems. Evidence from the literature

demonstrates that STEM activities significantly

enhance these abilities, thereby preparing students for

real-world challenges.

2.2 A Novel Design of STEM Activities:

Integrating STEM with

Introductory Iot Concepts and AI

Data Model Training Skills

Current STEM activities often lack integration with

contemporary technologies, limiting students’

exposure to practical applications. As technology

rapidly advances, incorporating new concepts into

STEM activities can enhance their appeal to students.

Integrating the IoT and AI into STEM education not

only can enhance learning by providing opportunities

for real-world applications but also can promote

digital creativity (Kong et al., 2024). The IoT is a

network of interconnected devices that gather data

and utilise embedded technology to make decisions

(Badshah et al., 2023). AI in IoT systems can

facilitate human–system interactions, making STEM

systems more flexible and interactive (Ghosh et al.,

2018). This integration helps students understand

complex technologies and apply them creatively,

thereby enhancing their problem-solving skills and

digital creativity. Therefore, there is a need for novel

STEM activities that incorporate the IoT and AI to

nurture digital creativity.

The proposed design addresses these gaps by

incorporating the IoT and AI into STEM activities,

providing students with hands-on experiences that

demystify complex systems design and encourage

exploration and creativity.

Evaluating a Teacher Development Course for Teaching STEM Activities with Introductory Internet of Things Concepts and AI Data Model

Training Skills Using the TPACK Framework: Problem-Solving and Digital Creativity

2.3 Six-Step Pedagogy: ‘to Play, to

Inquire, to Assemble, to Code,

to Create, to Reflect’

To effectively implement this novel design, we adopt

the six-step pedagogy – ‘to play, to inquire, to

assemble, to code, to create, to reflect’ (Kong, 2023).

This approach is designed to provide students with

hands-on experience in interacting with STEM

systems to stimulate their curiosity, develop their

problem-solving skills, and ultimately inspire their

digital creativity. The process begins with ‘to play’,

where students interact with the system and develop

an interest. This initial engagement naturally

progresses to ‘to inquire’, which fosters a deeper

understanding of the underlying mechanisms of the

system. Following this, ‘to assemble’ involves

students in the practical task of disassembling and

reassembling the system, which enhances their

understanding of the concepts of a STEM system. The

subsequent step, ‘to code’, integrates coding and

activities, particularly with AI data model training,

allowing students to apply their concepts in a

practical context. ‘To create’ empowers students to

share their ideas on building a new STEM system

using the technologies they have explored. Finally,

‘to reflect’ involves encouraging students to

consolidate their learning experiences and concepts

about IoT and AI data modelling skills. This

comprehensive pedagogy can be effectively

implemented in classroom settings, providing a

robust learning experience that significantly enhances

students’ problem-solving abilities and digital

creativity (Kong et al., 2024).

2.4 Guiding and Evaluating Teacher

Development with the TPACK

Framework

In this study, we utilise the TPACK framework

(Mishra & Koehler, 2006) to design a teacher

development course aimed at equipping teachers with

the essential knowledge and skills for implementing

STEM activities. The TPACK framework comprises

seven components, including three primary domains:

CK, TK, and PK. The interactions between these

domains create PCK, TCK, TPK, and TPACK. In the

STEM activities, CK pertains to introductory IoT

concepts, AI data model training skills integrated into

the STEM activities, and the problem-solving skills

required to develop the STEM systems. TK

encompasses general proficiency in using

technology, including computers, coding platforms,

microprocessors, and various electronic components.

PK involves instructional strategies for teaching the

STEM activities, such as guiding students in using

discussions and group activities to navigate the

problem-solving process and generate ideas for

digital creativity.

3 METHODOLOGIES

3.1 Participants and Procedure

Two hundred and one teachers from 108 primary

schools in Hong Kong joined the 6-hour teacher

development course. One hundred and ninety-one

teachers finished the pre-course and post-course

TPACK surveys, of whom 119 (63.30%) were male

and 72 (37.70%) were female. In addition, 110

(57.59%) of teachers taught mathematics, 28

(14.66%) taught general studies, 25 (13.09%) taught

English, 22 (11.52%) taught Chinese, and 6 (3.14%)

primarily taught information technology. Of these

teachers, 176 expressed their digital creativity ideas

in writing and sketches after the course. A course

evaluation form was also used to collect their views

on the course.

3.2 Instruments

3.2.1 Teachers’ TPACK Survey

In this study, we developed a survey tool for teachers

to self-assess their competency in delivering STEM

education fused with IoT concepts and AI

components within the TPACK framework. The

survey comprises 17 TPACK items, each rated on a

5-point Likert scale ranging from 1 (indicating strong

disagreement) to 5 (indicating strong agreement). All

17 items are listed in Appendix I of this paper. Our

assessment of the instrument’s reliability, conducted

using Cronbach’s alpha analysis, revealed strong

internal consistency, with values exceeding 0.85 for

both the pre-course survey (α = 0.97, N = 201) and

the post-course survey (α = 0.98, N = 191).

In addition to understanding how TPACK

pertains to the organization of learning and teaching,

it is crucial to explore methods for nurturing students'

digital creativity in solving real-life problems.

NACCCE (1999) distinguishes between teaching

creatively, which involves the teacher's own

creativity, and teaching for creativity, which focuses

on developing strategies that foster learners'

creativity. To effectively nurture students' creativity,

teachers must employ both approaches (Craft, 2005,

p. 44). Additionally, teachers must demonstrate their

CSEDU 2025 - 17th International Conference on Computer Supported Education

ability to cultivate ideas and transform knowledge

into solutions for real-life problems to promote

creative problem-solving using technology.

3.2.2 Written and Drawn Artefacts

Following the completion of the course, the primary

school teachers were invited to propose novel STEM

applications. These design artefacts were used to gain

insight into the advancement of the teachers’ digital

creativity after taking the course. The criteria for

evaluating the creativity of these artefacts are

presented in Table 1.

Table 1: Criteria for evaluating the creativity of the artefacts

proposed by the teachers after the course.

Criteria Mar

New STEM application designs incorporating

the introductory IoT concepts of sensing,

reasoning, and reacting

The clarity of the STEM application description

could be improve

Ideas that closely resembled the STEM

lications discussed in the course

3.2.3 Course Evaluation

The evaluation instrument for the professional

development course consists of 16 questions rated on

a 4-point Likert scale ranging from 1 = ‘strongly

disagree’ to 4 = ‘strongly agree’. This scale is

universally used for evaluating university courses.

The questionnaire items were designed to assess

various aspects of the course, such as its organisation,

its alignment with the course outline, its ability to

inspire and engage the participants, and the

effectiveness of the feedback and learning

opportunities provided. In addition to the Likert scale

items, the evaluation includes three short-answer

questions allowing participants to express their

feelings about the course. These questions ask the

participants to describe the most useful aspects of the

course and the reasons for their usefulness, to suggest

changes to help participants learn better and the

reasons for these suggestions, and to provide any

additional comments. This comprehensive evaluation

approach aims to capture both quantitative ratings and

qualitative feedback to understand participants’

experiences and identify areas for improvement.

3.3 The Teacher Development Course

The course comprised three teaching units designed

to incrementally deepen teachers’ understanding of

introductory IoT concepts and AI data model training

skills in STEM activities. They started by creating a

maze in a game featuring a character on a screen

controlled by a manually crafted physical joystick

linked to a microprocessor. Through interacting with

this system, the teachers developed a preliminary

understanding of the introductory IoT concepts of

sensing, reasoning, and reacting. The second unit

featured a ping-pong game (Kong et al., 2024). The

third unit introduced a smart face-changing game that

enables users to interact with the game using a prop

built with a microprocessor and a camera for AI

models in classifying images. Expanding upon the

previous unit’s learning, the teachers gained a deeper

understanding of AI model training. These processes

of data model training and subsequent discussions

fostered a foundational understanding of AI data

model training and practical application in the

development of IoT- and AI-integrated STEM

activities.

At the beginning of each unit, the teachers were

given an overview of the pertinent technologies and

content knowledge, including TK, CK, and TCK.

Subsequently, they were guided through a structured

teaching process comprising the six steps: ‘to play’

‘to inquire,’ ‘to assemble,’ ‘to code,’ ‘to create’, and

‘to reflect’. This approach was designed to provide

them with practical experience in teaching

methodologies and strategies for teaching problem-

solving skills and fostering digital creativity,

incorporating CK, PCK, TPK, and TPACK.

In the smart face-changing unit, the teachers were

initially introduced to the technological components

of the project, enabling them to acquire foundational

CK. Subsequently, they progressed through the six

steps of the pedagogical process. At the end of the

session, the teachers reflected on the IoT concepts, AI

data model training, and instructional design. They

also participated in collaborative idea-sharing

sessions, focusing on the innovation of novel systems

by leveraging their acquired knowledge, which

encompassed technological components and

experiences with training data models and utilising

them in the STEM context. This session was crucial

for fostering the teachers’ digital creativity and

empowering them with the skills to inspire the digital

creativity of their students.

This unit provides a prime illustration of the

pedagogical approach. During their interactions with

the system in the ‘to play’ stage, the teachers (1)

raised their eyebrows in front of the webcam to

trigger the Mask sprite, (2) shook the fan and gold

props to interact with the Gold sprite on the screen,

and (3) switched between the fan and gold props in

front of the webcams to control the Dress sprite and

Evaluating a Teacher Development Course for Teaching STEM Activities with Introductory Internet of Things Concepts and AI Data Model

Training Skills Using the TPACK Framework: Problem-Solving and Digital Creativity

background. Figure 1 depicts how IoT concepts

(sensing, reasoning, reacting) integrated with AI data

model training skills in the smart face-changing

system, aiding teachers in developing systems

thinking.

Figure 1: The IoT concepts of sensing, reasoning, and

reacting with AI data model training skills in the smart face-

changing system.

Subsequently, teachers then transitioned to the ‘to

inquire’ step, examining the underlying functionality

of the system through the concepts of sensing,

reasoning, and reacting.

Figure 2: The worksheet designed to support students in

critically reflecting on their interactions with the system.

A worksheet (Figure 2) was crafted to support the

students in critically reflecting on their interactions

with the system. This system comprises physical fan

and gold props with a microprocessor, battery box,

webcam, computer, and monitor. The built-in

accelerometer of the microprocessor detects

acceleration and sends data to the computer via

Bluetooth. The teachers were briefed on AI

confidence levels, which indicate the probability of

accurate decisions, highlighting that AI decisions

may not always be exact. By mapping the sensing and

reacting components, students gain a nuanced

understanding of how the system interprets inputs and

produces corresponding outputs. This exercise is

crucial for breaking down the development process

into smaller, more manageable tasks, thereby

enhancing students’ problem-solving skills (Figure

3).

Figure 3: The decomposition of tasks.

The reflective practice incorporated into this

worksheet is integral to effective instructional design,

aimed at reinforcing STEM computational thinking

concepts, particularly in the context of systems

engineering. As students engage with this content, they

will not only deepen their comprehension of the system

but also develop essential skills for future projects.

Prior to the ‘to assemble’ and ‘to code’ stages, the

teachers developed their abstract thinking through

exploration of interactions in the smart face-changing

project. They focused on (1) the pre-trained Face

Sensing model, (2) the microprocessor, and (3) the

data model trained using machine learning. This

understanding facilitated the task breakdown of a

problem into smaller tasks for incremental project

development.

Task 1 integrated the ‘Face Sensing’ extension

with the pre-trained AI data model (Figure 4).

Figure 4: Using pre-trained AI data model ‘Face Sensing’

extension.

CSEDU 2025 - 17th International Conference on Computer Supported Education

Task 2 focused on the steps of assembling the

props (as outlined in Figure 5) and coding (Figure 6).

The students had to cut cardboard according to a

template provided in the appendix of the student

guide. They then followed the instructions to

assemble the props and install the micro:bit and

battery case inside the cardboard props they created.

If the assembly sequence was not followed, the props

could not be built successfully. For example, if the

micro:bit was not placed within the cardboard before

the coloured cover was added, it could not be inserted

subsequently.

Figure 5: Prop assembly steps.

Figure 6: Blocks for controlling the Gold sprite based on

the interaction with the microcontroller.

In Task 3, the teachers trained a data model to

classify images as either a fan or gold. They then used

the confidence levels from the AI data model on MIT

RAISE Playground to control the costume changes of

the Dress sprite and the background based on images

captured by the webcam (see Figure 7).

Figure 7: Use of the AI data model for reasoning.

During the ‘to create’ activity, the teachers had to

brainstorm innovative solutions for solving real-

world problems. They had to explore possible AI data

models, sensors, and actuators to be used to solve a

new problem. Although constructing the proposed

artefacts was not required, this brainstorming session

enabled the teachers to apply their newly acquired

knowledge from the course.

At the end of each lesson, the teachers were

encouraged ‘to reflect’ on their learning experiences

and to consider ways to enhance the pedagogy of the

unit. This reflection focused on fostering their

abilities in solving-problem and inspiring their

students’ digital creativity with digital technologies

such as the IoT, AI, and physical objects.

Additionally, the teachers were guided to reflect on

their engineering systems thinking and to recognise

that encountering and overcoming failures is an

integral part of the learning process in STEM

activities.

4 RESULTS AND DISCUSSIONS

4.1 Teacher’s TPACK Development

A paired t-test was conducted using IBM SPSS

(Version 28) on 191 pairs of pre-course and post-

course survey responses to determine the significance

of the changes in teachers’ TPACK after completing

the teacher development course. The pre-course

means, post-course means, and t-test results for the

individual items are presented in Table 2.

Evaluating a Teacher Development Course for Teaching STEM Activities with Introductory Internet of Things Concepts and AI Data Model

Training Skills Using the TPACK Framework: Problem-Solving and Digital Creativity

Table 2: Paired t-test results of the teacher TPACK survey,

with each item scored on a 5-point Likert scale (N = 191).

Pre Post

Mean SD Mean SD t-value

CK 3.02 0.91 4.05 0.6 18.539***

TK 3.47 0.84 3.99 0.69 9.967***

PK 3.34 0.7 3.89 0.67 11.616***

PCK 3.35 0.78 3.91 0.72 10.398***

TCK 3.11 0.89 3.98 0.66 15.346***

TPK 3.07 0.91 3.94 0.71 14.299***

TPACK 3.16 0.83 3.91 0.69 13.604***

Overall 3.21 0.74 3.96 0.62 16.770***

Note. *: p < .05, **: p < .01, ***: p <. 001

The results indicate significant improvements

across all items, with medium to large effect sizes.

This suggests that the result is highly significant and

that there is strong evidence against the null

hypothesis. Overall, significant improvement was

observed (t(191) = 16.770, p < .001), with a large

effect size (Cohen’s d = 1.213). For Cohen’s d, a

value of 0.2 indicates a small effect, 0.5 a medium

effect, and 0.8 a large effect. Significant

improvements were found for all of the TPACK

items, with large effect sizes for CK (t(191) = 18.539,

p < .001, Cohen’s d = 1.341), PK (t(191) = 11.616, p

< .001, Cohen’s d = 0.841), TCK (t(191) = 15.346, p

< .001, Cohen’s d = 1.110), TPK (t(191) = 14.299, p

< .001, Cohen’s d = 1.035), and TPACK (t(191) =

13.604, p < .001, Cohen’s d = 0.984). Medium effect

sizes were observed for TK (t(191) = 9.967, p < .001,

Cohen’s d = 0.721) and PCK (t(191) = 10.398, p <

.001, Cohen’s d = 0.752).

The outcomes of the paired t-test analysis indicate

a significant enhancement of teachers’ confidence

levels concerning the instruction of AI and IoT

concepts, problem-solving proficiencies, and the

cultivation of digital creativity using appropriate

technological tools (TCK) because of the teacher

development programme. Furthermore, the course

elevated the teachers’ confidence in fundamental PK,

including collaborative learning techniques in STEM

education, as well as TPK, which encompasses the

effective utilisation of pedagogical strategies for

disseminating technological information. Moreover,

there was a marked improvement in the teachers’

confidence in TPACK, focusing on the delivery of

STEM content within the specialised framework of

STEM lessons. The findings also highlight

enhancements in teachers’ overall TK and their

capacity to apply CK to address students’ learning

challenges (PCK) after their participation in the

course.

4.2 Evaluation of the Digital Creativity

Development of the Teacher

Participants

4.2.1 Analysis of the Digital Creativity

Designs of the Teacher Participants

Two researchers independently assessed the teachers’

digital creative evaluation, achieving high inter-rater

reliability, with an ICC of 0.844 (p < .001) and 95%

confidence intervals between 0.797 and 0.880,

signifying substantial agreement. Of the 191 teachers

who completed the surveys, 176 had valid

submissions. Of these, 60 (34.09%) received 2 marks,

95 (53.98%) received 1 mark, and 21 (11.93%)

received 0 marks. The projects were then further

categorised based on the theme of the final

application: 54 (30.68%) in gaming, 44 (25.00%) in

health and fitness, 37 (21.02%) in teaching and

learning support, 12 (6.82%) in security, 10 (5.68%)

in support for learner diversity, 7 (3.98%) in smart

home/campus, 4 (2.27%) in environment monitoring,

4 (2.27%) in commercial use, and 4 (2.27%) in

inclusive society. Overall, 157 (82.20%) of the

teachers showcased their digital creativity by

designing IoT or IoT with AI systems following the

development course, and 139 (72.77%) expressed

their ideas about using AI.

4.2.2 Sample Designs from the Teacher

Participants

In this section, we present three instances of digital

creativity designs crafted by the participating

teachers.

The illustration in Figure 8 was done by a teacher

who devised a fitness training system aimed at

categorising the type of exercise being performed by

the user in front of a webcam.

In this drawing, the teacher demonstrated a

profound comprehension of engineering systems

thinking, clearly illustrating how individual

components interact to form a complete system. The

system utilises Teachable Machine to train a data

model for exercise identification. Images captured by

the webcam are processed by the application, which

then identifies the exercise type based on the trained

model. The teacher’s accompanying explanation

articulates the system’s logic: ‘If the player stand,

then the sprite stops. If run correctly, then move’ [‘If

the player stands still, then the sprite stops. If the

player runs in a correct pose and timing, then the

sprite also moves.’].

CSEDU 2025 - 17th International Conference on Computer Supported Education

Figure 8: A teacher’s conceptual design of a fitness training

programme proposing the use of an AI data model to

identify students’ physical activity.

To enhance user engagement and promote fitness,

additional features such as distance and time tracking

were incorporated. The teacher wrote, ‘Set a distance

or time limit to let player enjoy to play (keep fit)’.

Further development possibilities were also noted,

including the integration of a virtual opponent: ‘Can

add a computer run with the player, set different

level’. The teacher also suggested the addition of a

step-counting microprocessor, writing, ‘May add

micro:bit count steps’.

Figure 9 shows another teacher’s design of a

similar system for fitness but with more details about

its implementation. Within the realms of TK and

TCK, she explicitly demonstrated her ability to select

appropriate tools to underpin her concepts. This is

evidenced by her strategic utilisation of (1) the pre-

trained ‘Body Sensing’ model from Teachable

Machine for exercise classification and (2) the

accelerometer for tallying exercise repetitions and

displaying the variety of exercises completed.

Noteworthy is her adept use of variables to track the

different exercise types performed and exhibit them

on-screen, showcasing a sophisticated understanding

of technological applications.

Her design not only underscores her proficiency

in fundamental computational thinking (CT)

principles such as variables but also extends into the

domain of STEM CT concepts, encompassing

sensing, reasoning, and reacting and the application

of engineering systems thinking. The illustration not

only elucidates the screen layout but also delineates

the positioning and integration of the microprocessor,

indicating a comprehensive grasp of technological

integration and practical implementation strategies.

Figure 9: A teacher’s design of an AI fitness training

programme to identity students’ physical activities with

technological details.

Figure 10 shows a teacher’s design of an AI

energy saving system aimed at energy conservation,

using a reasoning approach grounded in the AI model.

The system is programmed to automatically switch

off the lights and other electrical appliances when a

room is unoccupied and to turn them on when the

room is occupied. The teacher has effectively

articulated the application of the AI model. Although

the teacher mentioned that a motor would be used to

control the lighting and other electrical appliances,

the explanation concerning the physical control

mechanisms of the lighting system was less detailed.

Figure 10: A teacher’s design of an AI energy saving

system for energy saving with proper reasoning based on

the AI model.

The professional development course was

designed to enhance educators’ knowledge and skills

in teaching STEM with IoT and AI within the

Evaluating a Teacher Development Course for Teaching STEM Activities with Introductory Internet of Things Concepts and AI Data Model

Training Skills Using the TPACK Framework: Problem-Solving and Digital Creativity

framework of CT education. This study used a

comprehensive evaluation methodology to

thoroughly analyse the course’s effectiveness and

identify areas for enhancement. Building on the

findings from the written and drawn artefacts

produced by the teachers, as presented in section 4.2,

the subsequent section (4.3) presents the findings

from the analysis of their feedback in the course

evaluation.

4.3 Evaluation

A course evaluation survey was conducted to gather

feedback from the participants immediately

following the completion of the 6-hour professional

development course. This survey, which used a

descriptive analysis approach, aimed to assess various

aspects of the course and identify areas for

improvement. The participants were asked to respond

to 16 items scored on a 4-point Likert scale ranging

from 1 = ‘strongly disagree’ to 4 = ‘strongly agree’ to

gauge their experiences and perceptions.

Additionally, three short-answer questions provided

opportunities for the participants to express their

feelings about the course, including the most useful

aspects, suggestions for improvement, and any other

comments. All 16 items are listed in Appendix II, and

the results are presented in Table 3.

Table 3: Descriptive analysis results of the teacher course

evaluation survey, with each item scored on a 4-point Likert

scale (N = 150).

Item Mean SD Item Mean SD

Q1 3.61 0.502 Q9 3.56 0.524

Q2 3.61 0.502 Q10 3.65 0.478

Q3 3.59 0.506 Q11 3.61 0.489

Q4 3.57 0.523 Q12 3.55 0.513

Q5 3.61 0.502 Q13 3.49 0.515

Q6 3.58 0.522 Q14 3.47 0.540

Q7 3.56 0.511 Q15 3.44 0.561

Q8 3.59 0.494 Q16 3.41 0.603

One hundred and fifty teachers completed and

submitted the evaluation form after attending the

teacher development workshop. The response rate

was 74.63%. The sixteen items were rated between

3.41 and 3.65, and the average was 3.56, indicating

that the teachers were satisfied with the quality of the

course. The course was highly regarded for its

organised delivery, achieving a mean score of 3.61

(SD = 0.502). The participants felt that the learning

and teaching were well-aligned with the course

outline, also scoring a mean of 3.61 (SD = 0.502). The

course inspired the participants to think and learn,

with a mean score of 3.59 (SD = 0.506). However,

addressing the participants’ specific learning needs

scored slightly lower, 3.57 (SD = 0.523). The course

effectively enhanced the participants’ course-related

knowledge or skills, with a mean of 3.61 (SD =

0.502). Providing appropriate feedback to enhance

learning was rated 3.58 (SD = 0.522), and

opportunities for learning from diverse sources

scored 3.56 (SD = 0.511). Guiding participants to

think from different perspectives achieved a mean

score of 3.59 (SD = 0.494), and encouraging

proactive engagement in learning received a mean

score of 3.56 (SD = 0.524). The instructors’

enthusiasm in teaching was highly appreciated, with

a score of 3.65 (SD = 0.478), and the overall teaching

quality was rated at 3.61 (SD = 0.489). The course’s

learning activities stimulated interest in teaching

STEM with IoT and AI in CT education, scoring 3.55

(SD = 0.513). The course also enhanced knowledge

in teaching STEM with IoT and AI and CT, scoring

3.49 (SD = 0.515), and understanding the pedagogy

of TPACK, with a score of 3.47 (SD = 0.540). The

participants felt that they acquired sufficient PK to

teach relevant STEM CT concepts, with a mean score

of 3.44 (SD = 0.561). However, confidence in

teaching and developing STEM activities through

Scratch programming was slightly lower, scoring a

mean of 3.41 (SD = 0.603).

These results indicate that the course was received

positively and was effective in enhancing the

participants’ knowledge and skills. However, there is

room for improvement in areas such as addressing

specific learning needs and building confidence in

using Scratch programming for STEM activities.

Following the quantitative assessment, the

teachers were asked to provide feedback on the most

useful aspects of this course and their reasons.

The feedback provided by the teachers reveals a

clear appreciation for various aspects of the course,

primarily revolving around the integration and

application of AI, hands-on activities, and teaching

methodologies.

The teachers appreciated the comprehensive

integration of AI in the curriculum, which enabled

them to bring real-world applications into their

teaching. Some of their feedback is as follows: ‘What

I appreciated most about this lesson was that it built

upon previous Scratch activities, deepened them by

incorporating AI, made it more fun, allowed us to

learn more, yet remained easy to execute’, ‘The demo

[enabled us to use] AI in our daily coding life’, ‘[I

learned] AI machine learning application’, ‘Theory

combined with coding is very practical’ and ‘[During

CSEDU 2025 - 17th International Conference on Computer Supported Education

the “to play” step,] Letting us to test and play the

products before teaching’.

The practical elements of the course, including

hands-on activities with micro:bit, Scratch, and

various sensors, were particularly valued. Some of the

teachers noted: ‘Having [the] opportunity to have

hands on experience was beneficial’, ‘[I learned] how

to use Teachable Machine in the project’, ‘Hands-on

experience is useful’, and ‘The use of Teachable

Machine in Scratch … is fun and not hard to use in

lessons’.

The teachers also emphasised the usefulness of

the pedagogical skills and methodologies imparted

during the course, as these would help them deliver

lessons more effectively and confidently. They said

that they valued the following: ‘to play’, which allows

us to get involved and draw our interests’, ‘Letting us

… test and play the products before teaching’, ‘The

pedagogical skills that help teachers deliver a more

confident lesson. Students will benefit from the

logical and systematic teaching’, ‘Giving ideas about

teaching AI’, ‘Teaching us how to deliver the

lessons’, and ‘Hands-on activities and handout

materials’.

The development course boosted the teachers’

confidence and inspired them to use technology tools

and pedagogy to deliver lessons and guide students in

more interesting ways. The teachers’ feedback

emphasises the importance of combining theoretical

knowledge with practical applications and engaging

teaching methods. Examples of the teachers’

responses with their related TPACK items are listed

in Appendix III of this paper.

5 CONCLUSION AND FUTURE

WORK

The findings from the teacher development course

highlight significant advancements in the teachers’

TPACK, particularly in areas such as CK, TCK, and

TPK. The substantial improvements across all

TPACK components, with medium to large effect

sizes, underscore the effectiveness of the course in

enhancing the teachers’ confidence and capabilities in

integrating IoT and AI into their teaching practices.

The evaluation of the digital creativity of the

teacher participants further demonstrates the

successful application of these technological

concepts. A significant majority of teachers

showcased their ability to design innovative smart

systems integrating IoT and AI, reflecting a deep

understanding of engineering systems thinking and

the practical application of IoT and AI principles. The

high inter-rater reliability in assessing these projects

confirms the robustness of the evaluation process.

The digital creativity evaluation provides further

insight into how the teachers applied their newly

acquired TK to develop TCK. The teachers

demonstrated advanced technological skills and

creativity in their projects, such as ideas on data

model training, using pre-trained models and sensors

to create interactive systems. The practical

application of TK and TCK in their design artefacts

highlights the significant impact of the course on their

ability to adopt and adapt new technologies in

educational contexts.

Additionally, the positive feedback from the

teacher development workshop, with high

satisfaction ratings and increased confidence reported

by the participants, reinforces the value of such

teacher development initiatives. The teachers

expressed that the course not only boosted their TK

but also inspired them to deliver lessons in more

engaging and effective ways.

In conclusion, the teacher development course

significantly enhanced the teachers’ competency in

teaching novice IoT- and AI-integrated STEM

activities and thus boosted their digital creativity.

However, it was limited by a lack of evidence related

to primary students or the potential changes in

teachers’ confidence after real classroom practice.

Future work should focus on evaluating the impact on

students’ understanding and problem-solving abilities

and on comparing teachers’ post-course and in-

practice competencies. This will help determine the

long-term effectiveness of the course in real

educational settings.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the funding

support for this CoolThink@JC project from the

Hong Kong Jockey Club Charities Trust (Project No.

EdUHK C1136).

REFERENCES

Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari,

M., & Ayyash, M. (2015). Internet of Things: A survey

on enabling technologies, protocols, and applications.

IEEE Communications Surveys Tutorials, 17(4), 2347–

2376. https://doi.org/10.1109/COMST.2015.2444095.

Badshah, A., Ghani, A., Daud, A., Jalal, A., Bilal, M., &

Crowcroft, J. (2023). Towards smart education through

Evaluating a Teacher Development Course for Teaching STEM Activities with Introductory Internet of Things Concepts and AI Data Model

Training Skills Using the TPACK Framework: Problem-Solving and Digital Creativity

Internet of Things: A survey. ACM Computing Surveys,

56(2), 1–33.

Cavlazoglu, B., & Stuessy, C. (2017). Changes in science

teachers’ conceptions and connections of STEM

concepts and earthquake engineering. The Journal of

Educational Research, 110(3), 239–254.

Craft, A. (2005). Creativity in schools: Tensions and

dilemmas. Routledge.

Ghosh, A., Chakraborty, D., & Law, A. (2018). Artificial

intelligence in Internet of Things. CAAI Transactions

on Intelligence Technology, 3(4), 208–218.

https://doi.org/10.1049/trit.2018.1008.

Hanif, S., Wijaya, A. F. C., & Winarno, N. (2019).

Enhancing students’ creativity through STEM project-

based learning. Journal of Science Learning, 2(2), 50.

https://doi.org/10.17509/jsl.v2i2.13271.

Hsu, T. C., Hsu, T. P., & Lin, Y. T. (2023, March). The

artificial intelligence learning anxiety and self-efficacy

of in-service teachers taking AI training courses. 2023

International Conference on Artificial Intelligence and

Education (ICAIE) (pp. 97–101). IEEE.

https://doi.org/10.1109/ICAIE56796.2023.00034.

Kim, S., Jang, Y., Kim, W., Choi, S., Jung, H., Kim, S., &

Kim, H. (2021). Why and what to teach: AI curriculum

for elementary school. Proceedings of the AAAI

Conference on Artificial Intelligence, 35(17), 15569–

15576. https://doi.org/10.1609/aaai.v35i17.17833.

Kong, S. C. (2023). Pedagogical design of STEM activities

for developing problem-solving skills and digital

creativity of primary students in the Internet of Things

era: Six-step STEM pedagogy. In W. M. So, Z. H. Wan,

& T. Luo (Eds.), Cross-disciplinary STEM learning for

Asian primary students: Design, practices and

outcomes (pp. 147–163). Routledge.

Kong, S. C., & Yang, Y. (2024). A human-centered

learning and teaching framework using generative

artificial intelligence for self-regulated learning

development through domain knowledge learning in K-

12 settings. IEEE Transactions on Learning

Technologies, 17, 1562–1573. https://doi.org/10.

1109/TLT.2024.3392830.

Kong, S. C., Yang, Y., & Yeung, W. K. (2024). A proposed

TPACK model of teaching STEM with AI components:

Evaluating a teacher development course for fostering

digital creativity. Proceedings of the 16th International

Conference on Computer Supported Education –

Volume 2 (pp. 163–170).

Lo, C. K. (2021). Design principles for effective teacher

professional development in integrated STEM

education: A systematic review. Educational

Technology & Society, 24(4), 136–152.

https://doi.org/10.30191/ETS.202110_24(4).0011.

Mishra, P., & Koehler, M. J. (2006). Technological

pedagogical content knowledge: A framework for

teacher knowledge.

Teachers College Record, 108(6),

1017–1054.https://doi.org/10.1111/j.1467-

9620.2006.00684.x.

National Advisory Committee on Creative and Cultural

Education (NACCCE). (1999). All our futures:

Creativity, culture and education. Report to the

Secretary of State for Education and Employment.

Sudbury: Department for Education and Employment.

Sullivan, F. R., & Heffernan, J. (2016). Robotic

construction kits as computational manipulatives for

learning in the STEM disciplines. Journal of Research

on Technology in Education, 48(2), 1–24.

https://doi.org/10.1080/15391523.2016.1146563.

Touretzky, D., Gardner-McCune, C., Martin, F., &

Seehorn, D. (2019). Envisioning AI for K–12: What

should every child know about AI? Proceedings of the

Thirty-Third AAAI Conference on Artificial Intelligence

(AAAI-19) (pp. 9795–9799). AAAI Press.

https://doi.org/10.1609/aaai.v33i01.33019795.

Wang, Y. Y., & Wang, Y. S. (2022). Development and

validation of an artificial intelligence anxiety scale: An

initial application in predicting motivated learning

behavior. Interactive Learning Environments, 30(4),

619-634. https://doi.org/10.1080/10494820.2019.

1674887.

APPENDIX I

Number Item

Q1 I understand ‘sensing–reasoning–reacting’ in the

operation process of IoT and related concepts.

Q2 I have sufficient knowledge about STEM

education in the IoT era.

Q3 I can use computational thinking practices, such as

sequencing, conditional reasoning, causal

reasoning, and engineering systems thinking, for

problem-solving in STEM activities.

Q4 I can learn new technology easily.

Q5 I can solve technical problems when using

technology.

Q6 I can adapt my teaching based upon what students

currently understand or do not understand.

Q7 I usually conduct student learning activities in a

collaborative way.

Q8 I can design some learning activities for students

to develop problem-solving skills.

Q9 I am able and willing to provide a complete STEM

activity artefact for my students to play and to

inquire.

Q10 I can identify and handle what learning difficulties

students might have on technological innovation in

STEM education.

Q11 I understand the functions that sensors,

microprocessors, and actuators perform in IoT

systems.

Q12 I believe that the electronic parts of STEM

teaching tools (e.g., micro:bit, M5Stick), such as

sensors, microprocessors, and actuators, can be

used to foster students’ digital creativity.

Q13 I can use STEM tools (e.g., micro:bit, M5Stick) to

organise STEM activities and foster students’

digital creativity.

CSEDU 2025 - 17th International Conference on Computer Supported Education

Q14 I can use appropriate teaching methods to teach

students to inquire and understand various

electronic parts (e.g., sensors, actuators) used in

STEM activities.

Q15 I can teach STEM lessons that appropriately

combine the content of STEM, technological

innovation, and proper teaching approaches.

Q16 I can select and use technologies in my classroom

that enhance what I teach, how I teach, and what

students learn.

Q17 I can provide support and leadership in helping

others to coordinate the use of STEM education,

technological innovation, and teaching approaches

at my school and/or district.

APPENDIX II

Number Item

Q1 The course was delivered in an organised way.

Q2 The learning and teaching aligned with the course

outline.

Q3 Participants were inspired to think and learn.

Q4 Participants’ needs in learning were addressed.

Q5 Participants’ course-related knowledge or skills

were enhanced.

Q6 Appropriate feedback to enhance learning was

provided.

Q7 Opportunities were provided for the participants to

learn from a variety of sources or methods.

Q8 Participants were guided to think from different

perspectives.

Q9 Participants were encouraged to proactively

engage in their own learning.

Q10 The teacher of this course was enthusiastic about

teaching

Q11 The overall teaching was of high quality.

Q12 The learning activities of the course stimulated my

interest in the understanding of teaching STEM

with IoT and AI (artificial intelligence) in

computational thinking education.

Q13 The course enhanced my knowledge of how to

teach STEM with IoT and AI (artificial

intelligence) and computational thinking.

Q14 The course enhanced my knowledge of the

pedagogy of TPACK in teaching STEM with IoT

and AI (artificial intelligence) and computational

thinking education.

Q15 I have acquired sufficient knowledge in pedagogy

to teach relevant STEM computational thinking

concepts, practices and perspectives.

Q16 I am confident in teaching and developing STEM

activities through Scratch programming.

APPENDIX III

Selected Quotes from Teachers TPAC

- ‘Different sensing techniques were integrated into

coding.’

- ‘The use of the data model extracted from the

teachable machine and use it in an application in

Scratch. It is fun and easy to implement in lessons.’

- ‘Showing how we can make good use of different

Scratch extensions to integrate micro:bit and AI to

make simple games’

- ‘Teaching how to play the sensor and AI. It is fun

and useful.’

- ‘Learning [I learned] the integration of STEM,

AI, and coding.’

CK,

TPK,

TPACK

- ‘The teaching materials are abundant and can be

effectively used in the classroom.’

- ‘Providing suitable material and hands-on

practice.’

- ‘Learning [I learned] the use of AI data model

training and knowing its impact and use in STEM

activities and learn the methodology to teach

students.’

TK,

TPK,

CK,

PCK,

TPACK

- ‘[I learned] how to use the trained data in raise

MIT edu [MIT RAISE Playground platform].’

- ‘[I learned] the use of teachable machine in

Scratch. It is fun and not hard to use in lessons.’

- ‘Showing how to mix the usage of Scratch and

micro:bi

.’

TK,

TPK

- ‘The pedagogical skills help teachers to deliver

STEM lessons with confidence. Students will

benefit from the logical and systematic teaching.’

- ‘[It] encouraged us to encourage students to create

new items based on what we have learnt from the

given material.’

- ‘It gave ideas about teaching AI.’

- ‘It provided teaching with examples that can be

applied in the classroom.’

- ‘It let us test and play with the products before

teaching.’

PCK

Evaluating a Teacher Development Course for Teaching STEM Activities with Introductory Internet of Things Concepts and AI Data Model

Training Skills Using the TPACK Framework: Problem-Solving and Digital Creativity