Education ยท Psychological Research
Research on Improving the AI Teaching Self-efficacy of Computer Science Teachers
Sun Junmeia, Yang Yanb, Ma Honglianga
(Shaanxi Normal University a. School of Education; b. School of Journalism and Communication, Xi’an, Shaanxi 710061)
Abstract: At present, the construction of the artificial intelligence curriculum in K-12 schools has been included in the national strategic development plan. However, the shortage of teachers has become the bottleneck of curriculum construction. Therefore, it is necessary to vigorously implement effective teacher professional development. Promoting teaching self-efficacy will affect teachers’ teaching engagement, teaching practice and students’ learning effect. Hence this study designed a professional development knowledge system based on TPACK theory, constructed a blended professional development combing with ways to improve teaching self-efficacy and effective teachers’ professional development elements, and carried out a 25-days intervention with 40 potential AI teachers as participants. In terms of improving teaching self-efficacy, the data analysis results of the questionnaire and interview showed that this kind of teacher professional development significantly improved CS teachers’ AI teaching self-efficacy, and caused no significant differences in CS teachers with different prior knowledge. These findings reveal the effectiveness of this professional development model in improving the AI teaching self-efficacy of K-12 CS teachers and may have a positive impact on their teaching practice and students’ learning results.
Keywords: computer science teacher; AI teaching self-efficacy; teacher professional development; TPACK
1. Current Status of Research on Cultivating Teaching Self-efficacy of Computer Science Teachers
(1) Teaching Self-efficacy and Cultivation Pathways
Self-efficacy is the strength of an individual’s belief in their ability to complete tasks, consisting of two elements: outcome expectation beliefs and efficacy beliefs[1]. Riggs and Enochs applied Bandura’s theory to the study of teaching self-efficacy, defining it from a bidirectional perspective, believing that teaching self-efficacy is the extent to which teachers believe they can have a positive impact on students’ learning, including teaching efficacy beliefs and outcome expectations[2].
Research indicates that teaching self-efficacy can be cultivated through four pathways: mastery experiences, vicarious experiences, social persuasion, and physiological arousal[3-5]. Mastery experiences refer to the successful experiences of completing specific tasks[3]. Compared to simple success, overcoming difficulties to gain mastery experiences can cultivate a resilient form of self-efficacy, which is more beneficial for enhancing teachers’ self-efficacy[6]. Vicarious experiences refer to the process of gaining self-efficacy through observing similar successful social models, where the similarity and relevance of the social model to oneself are considered key factors in acquiring self-efficacy through vicarious experiences[5,7]. Social persuasion refers to the verbal interactions teachers receive from other important figures in the teaching environment regarding their performance. In terms of social persuasion, emphasis should be placed on positive evaluations of teachers’ self-improvement rather than peer comparisons. Physiological arousal refers to the perception of an individual’s emotional state and physiological responses[5], where psychological and emotional arousal can also increase perceptions of capability or incapacity.
(2) Forms of Professional Development for Computer Science Teachers
Computer science teachers are a group that has grown alongside the development of information technology education. Based on different activity contexts, professional development activities for computer science teachers can be divided into face-to-face, online, and blended professional development activities. Traditional forms of professional development for computer science teachers include face-to-face workshops and summer learning classes[8]. These professional development activities typically include short-term lectures, work examples, open projects, and contextualized information technology teaching practices[9]. With the advent of the intelligent era, researchers have also utilized online learning for teacher professional development activities, allowing teachers to access community-centered MOOCs at any time and from anywhere[10], or use email[11] and online practice communities for learning[12]. For example, Goode et al. transformed face-to-face professional development activities into online formats to promote the professional development of computer science teachers from different regions, enabling them to be competent in new, equity-focused computer curriculum teaching practices[13].
Currently, blended professional development that combines face-to-face workshops with online learning activities such as webinars and community meetings is increasingly demonstrating its value in the professional development of computer science teachers. For instance, based on Rogers’ blended learning theory, Mashikhi and Soliman designed a training program for 40 elementary school computer teachers and verified the effectiveness of the training project[14]. Other research has utilized the Moodle platform to conduct blended training for K-12 teachers on Scratch, finding that this model promotes teachers’ computational thinking development[15]. Additionally, some studies have supported collaboration and communication among teachers through online professional development communities after formal face-to-face workshops and courses, allowing teachers to continuously receive peer support[16].
2. Research Questions
In 2017, the State Council issued the “New Generation Artificial Intelligence Development Plan,” clearly proposing the implementation of a national intelligent education project, establishing artificial intelligence-related courses at the K-12 level[17]. In recent years, the introduction of artificial intelligence courses in K-12 schools in some regions of China has gradually advanced. From the perspective of practical application and implementation, K-12 artificial intelligence education is currently mainly concentrated in information technology courses, STEM courses, and maker courses[18]. Most teachers of artificial intelligence-related courses are concurrently serving as existing information technology teachers. This clearly poses new requirements for the AI knowledge, abilities, and teaching of K-12 information technology teachers. However, the country has not yet established a teacher training program for artificial intelligence that serves K-12 schools, nor a corresponding teacher qualification certification system, leading to a lack of professional talents who understand both artificial intelligence technology and education in higher education institutions[19]. Moreover, professional training in artificial intelligence education has not been systematically implemented, and the professional knowledge reserves of information technology teachers still fall short of the requirements for teaching artificial intelligence courses, necessitating relevant professional development to enhance their professional capabilities in AI teaching.
In terms of teacher professional competency, teaching self-efficacy is a potential structure that influences teachers’ content, organization of teaching activities, and teaching behaviors[20] and has a significant positive impact on teachers’ teaching thinking[21] and a strong correlation with students’ learning outcomes[22]. Therefore, improving the AI teaching self-efficacy of information technology teachers not only positively impacts their AI teaching thinking and practices but also has a beneficial effect on students’ AI learning. Additionally, Desimone[23] suggested that teachers participating in effective professional development activities can promote the development of teachers’ knowledge and self-efficacy, thereby facilitating changes in teaching practices and ultimately improving students’ learning levels. Through literature research, current empirical studies on the self-efficacy of information technology teachers focus on two aspects: one is conducting surveys on the current situation of self-efficacy[24], influencing factors[25-26], etc., while the other focuses on professional development activities to enhance teachers’ information technology teaching self-efficacy[27], programming teaching self-efficacy[28], computational thinking teaching self-efficacy[29], and the self-efficacy of integrating computational thinking into teaching[30]. It is evident that neither aspect of research has addressed the AI teaching self-efficacy of information technology teachers.
In summary, this paper aims to provide professional development activities for K-12 information technology teachers to enhance their AI teaching self-efficacy. The specific research content includes: (1) designing a professional development content system and activity framework to enhance the AI teaching self-efficacy of K-12 information technology teachers; (2) verifying the effectiveness of this professional development activity in improving the AI teaching self-efficacy of K-12 information technology teachers; (3) examining whether there are differences in the effectiveness of this professional development activity in enhancing AI teaching self-efficacy among information technology teachers with different prior knowledge.
3. Theoretical Framework for Teacher Professional Development
(1) Knowledge System Design
American scholars Koehler and Mishra proposed the Technological Pedagogical Content Knowledge (TPACK) framework based on Pedagogical Content Knowledge (PCK), which includes three core elements: Content Knowledge (CK), Pedagogical Knowledge (PK), and Technological Knowledge (TK); as well as four composite elements: PCK, Technological Content Knowledge (TCK), Technological Pedagogical Knowledge (TPK), and TPACK[31]. Since then, numerous theoretical and practical studies on TPACK have been conducted by scholars both domestically and internationally.
Currently, there are two interpretations of artificial intelligence knowledge: the first refers to AI technology knowledge that empowers education and teaching; the second refers to the artificial intelligence knowledge that is part of K-12 subject curriculum content. This professional development for AI teachers aims to improve the teaching self-efficacy of information technology teachers, thus the AI knowledge as part of K-12 curriculum content is the core knowledge of this professional development. Based on TPACK theory, this paper constructs a TPACKAI framework for professional development (as shown in Table 1), including CKAI, TPKAI, TCKAI, PCKAI, and TPACKAI. CKAI refers to knowledge about artificial intelligence, including the “five major concepts of artificial intelligence”[32] and AI applications; TCKAI pertains to graphical programming related to artificial intelligence, including using software and hardware to learn AI; PCKAI involves teaching methods and instructional design related to artificial intelligence; TPKAI focuses on AI teaching tools, including relevant applications, websites, and tools; TPACKAI emphasizes the integration of programming, technology, and pedagogy to teach AI, such as developing school-based AI teaching materials.
Table 1 Knowledge System for Enhancing AI Teaching Self-efficacy of Information Technology Teachers Based on TPACK
(2) Activity Framework Design
This paper constructs a professional development activity framework aimed at enhancing AI teaching self-efficacy (as shown in Figure 1). In terms of activity format, a blended professional development activity combining online and offline elements is designed, with online activities including online course learning, programming practice, and community discussions; offline activities encompass practical community building, expert lectures, hands-on practice, observational learning, and teaching practice workshops in various forms. In terms of specific activity design, based on the characteristics of information technology subject teachers and AI knowledge, corresponding professional development activities are designed in conjunction with effective teacher professional development design elements proposed by Darling-Hammond et al.[33] (content focus, active learning, supportive collaboration, practical templates, and duration) and the three cultivation pathways for teachers’ teaching self-efficacy (mastery experiences, vicarious experiences, and social persuasion).
Figure 1 Professional Development Activity Framework for Enhancing AI Teaching Self-efficacy of Information Technology Teachers
In terms of content focus, the activity design uses the TPACKAI framework to construct a professional development knowledge system, focusing on AI curriculum knowledge, technology, and pedagogy. In terms of active learning and collaboration, by incorporating project-based teaching, cooperative learning, and creative learning into professional development activities, a series of activities such as group discussions, project creation, and sharing insights are conducted, allowing teachers to share ideas and collaborate in professional learning, including collaboratively designing programming projects, discussing instructional design, and organizing peer reviews. In terms of expert demonstrations and teaching showcases, high-quality teaching demonstration materials for AI classes are provided to teachers, including examples of AI programming projects, instructional designs, and teaching observations. In terms of duration, the activities consist of a 10-day, 60-hour intensive face-to-face training and a 15-day, 36-hour online course learning phase.
4. Research Design
(1) Research Subjects
The subjects of this study are 40 K-12 information technology teachers from various cities in Shaanxi Province, including 22 males and 18 females. Preliminary survey data show that among the 40 participating teachers, 42.11% had previously received training related to artificial intelligence, 28.95% taught AI-related courses, 28.95% had attended AI education-related lectures at their schools, 44.74% had organized students to participate in AI-related competitions, and 18.42% had taught content related to artificial intelligence.
(2) Research Method
This paper adopts a quasi-experimental single-group pre-test and post-test design, utilizing questionnaire and interview methods to collect experimental data, as shown in Table 2. Before the face-to-face professional development for teachers begins, teachers are organized to participate in a pre-test to assess their AI knowledge and teaching self-efficacy. The first stage of the face-to-face professional development lasts for 10 days, and the second stage of online professional development lasts for 15 days, after which a scale is used to test their AI teaching self-efficacy, and semi-structured interviews are conducted with selected teachers.
(3) Research Tools
1. AI Teaching Self-efficacy Scale
The AI teaching self-efficacy scale is adapted from the science teacher self-efficacy scale developed by Riggs and Enochs, which is based on the dimensions of teaching outcome efficacy and teaching behavior efficacy, containing 25 items, using a 5-point Likert scale[2]. The author localized the original scale and modified some items, resulting in the AI teaching self-efficacy scale, which contains 10 items, including 4 items on teaching outcome expectations and 6 items on teaching efficacy beliefs (as shown in Table 3). Subsequently, the scale’s reliability and validity were tested among 111 K-12 information technology teachers, yielding a KMO value of 0.791>0.7, indicating a certain correlation between items; the rotated factor loading matrix confirmed that the scale has good convergent validity; the Cronbach’s Alpha coefficient for teaching self-efficacy is 0.822>0.7 (teaching efficacy belief Cronbach’s Alpha=0.787>0.7, teaching outcome expectation Cronbach’s Alpha=0.738>0.7), indicating that the scale has a certain degree of stability and consistency. A total of 38 questionnaires were collected in the pre-test, with a recovery rate of 95%; 34 questionnaires were collected in the post-test, with a recovery rate of 85%. Due to missing data (some participants completed the pre-test but did not respond to the post-test), effective questionnaires were screened using the pre-post matching method (only participants with matched pre- and post-tests were included in the analysis) to maximize sample utilization[34]. After excluding invalid questionnaires, a total of 32 valid questionnaires remained, with a valid questionnaire rate of 80%.
Table 2 Experimental Design
Table 3 AI Teaching Self-efficacy Scale and Rotated Factor Loading Matrix
2. AI Knowledge Test Questionnaire
The K-12 AI education initiative (AI4K12) divides the AI knowledge required at the K-12 stage into five themes: perception, representation and reasoning, machine learning, human-computer interaction, and social impact[32]. Based on these five themes, this study designed an AI knowledge test questionnaire with a total of 15 items (10 multiple-choice questions and 5 multiple-answer questions). Each question has four options, with correct answers scoring 1 point for multiple-choice questions and 2 points for multiple-answer questions (missing options score 1 point, and incorrect answers score 0 points). Statistical analysis shows that the discrimination index for each question in this test questionnaire ranges from 0.32 to 0.41, indicating good discrimination and a certain degree of discernibility; the average difficulty index for the entire test is 0.48, indicating moderate difficulty.
3. Semi-structured Interview Outline and Selection of Interviewees
For the teaching self-efficacy of AI course teachers, this study designed a semi-structured interview outline, including: (1) After this training, do you feel more confident in conducting relevant AI courses/activities/content? Why? (2) Do you feel more confident in guiding students? Why? (3) Will you conduct AI-related courses/activities/content in your school in the future? Why? (4) What difficulties do you think may arise in conducting relevant course activities/themes/content?
In selecting interviewees, for the 32 valid pre-test questionnaires, the S-W test was used to check the normal distribution of the data, with results showing that the pre-test data for teachers’ teaching self-efficacy (W= 0.966, p= 0.406>0.05) and AI knowledge (W= 0.964, p= 0.359>0.05) both followed a normal distribution, indicating that the median and mean values are very close. Therefore, either the median or mean can be used as the reference line to segment the scatter plot formed by the two dimensions of teachers’ teaching self-efficacy and AI knowledge into four quadrants (as shown in Figure 2), thus categorizing participating teachers into four types: T1 (high teaching self-efficacy, low AI knowledge), T2 (high teaching self-efficacy, high AI knowledge), T3 (low teaching self-efficacy, low AI knowledge), and T4 (low teaching self-efficacy, high AI knowledge). One representative teacher from each category was selected for semi-structured interviews.
Figure 2 Teacher Classification Quadrant Diagram
5. Research Results
(1) Impact on AI Teaching Self-efficacy of Computer Science Teachers
Since the sample size is less than 50, this study used the S-W test to check the normality of the change scores, with results showing that the change scores for AI teaching outcome expectations (W= 0.965, p= 0.378>0.05), AI teaching efficacy beliefs (W= 0.946, p= 0.109>0.05), and teaching self-efficacy (W=0.972, p=0.561>0.05) all followed a normal distribution. Therefore, paired sample t-tests were used to analyze the impact of the teacher professional development program on the AI teaching self-efficacy of information technology teachers, with data analysis results shown in Table 4.
From the results in Table 4, it can be seen that teachers who participated in this AI training program showed a significant overall improvement in teaching self-efficacy; moreover, both sub-dimensions of teaching efficacy beliefs and teaching outcome expectations also showed significant improvement. For instance, regarding teaching efficacy beliefs, T2 mentioned in the interview: “I will adjust the curriculum settings after returning, adding some content related to AI, maybe one or two classes per semester or conducting some relevant activities to introduce AI and graphical programming to students.” In terms of teaching outcome expectations, T4 said: “After returning, I will apply for Arduino boards to experiment with some projects myself. I have a strong desire to complete them and guide students to create; if students have creative ideas, they can incorporate those ideas into their projects.”
(2) Impact on AI Teaching Self-efficacy of Computer Science Teachers with Different Prior Knowledge
Before the professional development activities, the AI knowledge questionnaire was used to test the participating teachers’ knowledge reserves in the AI-related field. Based on the test results, the participating teachers were divided into low, medium, and high groups according to their prior knowledge. To explore whether there are significant differences in the impact on AI teaching self-efficacy among computer science teachers with different prior knowledge, this study conducted one-way ANOVA to determine if significant differences exist, with data analysis results shown in Table 5.
Table 4 Comparison of AI Teaching Self-efficacy Pre-test and Post-test (N=32)
Note: ** Significant correlation at the 0.05 level (two-tailed).
Table 5 One-way ANOVA for AI Teaching Self-efficacy of Computer Science Teachers with Different Prior Knowledge
The results in Table 5 indicate that there are no significant differences in the improvement of AI teaching self-efficacy and its two sub-dimensions among computer science teachers with different prior knowledge. For example, T1, who had relatively low prior knowledge, stated: “I had never been exposed to graphical programming software before, but now I feel that regardless of whether I can program myself, I can understand it, and I no longer worry about this content. Furthermore, I didn’t know what artificial intelligence was before, but now I have a general understanding, and I feel more confident in my teaching.” T3, who had moderate prior knowledge, said: “Our school implemented related courses relatively early, using text-based programming software, which was quite challenging for us. After this training, I plan to switch to visual programming software so that students can truly enjoy realizing their ideas through programming.” T4, who had high prior knowledge, mentioned: “Our school has programming clubs using software like Scratch and Coding Cat. Previously, I only taught students how to program, but due to a lack of equipment, I had not engaged in hands-on operations related to hardware. After this training, I have developed a strong interest in hands-on operations, and I plan to apply for and purchase many Arduino boards.”
6. Conclusion and Discussion
This study designed a professional development knowledge system aimed at enhancing the AI teaching self-efficacy of K-12 information technology teachers based on TPACK theory, and constructed a blended professional development activity framework combining pathways to enhance teaching self-efficacy and effective teacher professional development design elements. After a 25-day implementation of the first round, the research found that the AI teaching self-efficacy of information technology teachers significantly improved at the overall level, as well as in the two sub-dimensions of teaching efficacy beliefs and teaching outcome expectations; and there were no significant differences among information technology teachers with different prior knowledge. These findings reveal the effectiveness of the professional development knowledge system and activity framework constructed in this research in enhancing the AI teaching self-efficacy of K-12 information technology teachers. Therefore, professional development activities aimed at enhancing AI teaching self-efficacy of information technology teachers should possess the following two characteristics.
(1) Effectiveness of Training Content Design in Enhancing AI Teaching Self-efficacy of Information Technology Teachers
To enhance the AI teaching competency of K-12 information technology teachers, a scientific, reasonable, and effective professional development content system needs to be constructed. Unlike PCK, TPACK is an effective knowledge framework for integrating technology into teacher professional development, applicable to guiding training content design for teachers across various subjects in technology-integrated teaching[35,36]. For the specific technology of artificial intelligence, there are two types of TPACK frameworks: one is the TPACK framework for teachers of various subjects applying AI technology in subject teaching, such as the AI-TPACK framework constructed by Yan Zhiming et al.[37] and the ethical knowledge framework (AIPCEK) driven by “AI + subject teaching” constructed by Deng Guomin et al.[38]; the other is the TPACK framework (TPACKAI) for information technology teachers conducting artificial intelligence course teaching. These two frameworks have significant differences in teacher professional development, mainly differing in training subjects and goals, leading to notable differences in training content frameworks. Existing research has mostly focused on the first framework and application level, emphasizing teachers’ abilities to integrate AI technology into information-based teaching. Technology is crucial for the TPACK framework, but three types of technology need to be distinguished: technology as a specific subject tool (e.g., TCK), technology as a learning tool, and conventional ICT tools used to support professional development or practice communities[39]. In this study, all three types of technology were utilized, particularly graphical programming technology and open-source hardware technology for learning AI knowledge. It is precisely because these operable and verifiable TCK technologies were used that teachers found learning AI knowledge intuitive, understandable, and effective.
The professional development project in this study targeted information technology teachers and did not adopt a purely AI skills-oriented training content design. Instead, it aimed at enhancing AI teaching competency, constructing a professional development content system centered on TPACKAI, integrating CKAI, TPKAI, TCKAI, PCKAI, and TPACKAI to comprehensively enhance information technology teachers’ knowledge of AI, technology-integrated AI knowledge, AI teaching knowledge, and technology-integrated AI subject teaching methods. The research results indicate that this teacher professional development project can effectively improve the AI knowledge of information technology teachers, as well as their teaching abilities and self-efficacy. Thus, the teacher professional development knowledge framework based on TPACKAI can achieve integration and balance between AI subject knowledge, pedagogy, and relevant supporting technologies (graphical programming, open-source hardware, etc.), helping teachers with different prior knowledge reserves to form flexible and contextualized understandings and cognitions of AI curriculum knowledge, thereby positively impacting their AI teaching competency.
(2) Effectiveness of Activity Framework Design in Enhancing AI Teaching Self-efficacy of Information Technology Teachers
Having an effective knowledge framework alone is insufficient to ensure the comprehensive enhancement of information technology teachers’ AI teaching competency; it also requires diverse activity formats to support the effective implementation of training content. Essentially, knowledge content and activity formats are closely related and complementary. This study integrates various activity formats that reflect effective teacher professional development elements and enhance teaching self-efficacy, particularly including hardware and software-based programming practices and operations, observational learning at AI education demonstration schools, and group teaching discussions and sharing in workshop formats.
Due to the singularity of activity formats, teacher professional development activities have long formed a top-down transmission modelโ”lecture-style” professional development. Unlike purely lecture-style training and distinct from “fragmented, superficial” thematic discussions[40], the activities designed in this study that embody effective teacher professional development elements such as content focus, active learning, and collaboration possess characteristics that integrate theoretical learning with practical operations, programming skills with instructional design, and independent learning with cooperative learning, thus effectively stimulating teachers’ enthusiasm and initiative for learning, better supporting knowledge social construction, practical experience exchange, and transformation of teaching practices among teachers. First, the course content of professional development activities is focused, and the knowledge system design encompasses not only subject-specific knowledge but also corresponding technology and pedagogy studies to promote student learning[41,42]. Second, during participation in professional development projects, teachers are active participants rather than passive recipients of knowledge. The practical sections of the professional development activities, such as operating open-source hardware devices and programming, discussing instructional designs in teaching practice workshops, and independent learning and programming in online learning sections, provide opportunities for teachers to gain mastery experiences and self-directed learning. Third, professional development activities provide opportunities for collaboration with peers, requiring teachers to share their thoughts and receive feedback from peers. During expert lectures and school-enterprise visits, participating teachers, as learners and observers, gain suitable teaching experiences for conducting AI education, which is one effective way to obtain vicarious experiences[13]. Meanwhile, teachers with similar teaching backgrounds, years of experience, and teaching levels can strengthen the influence of social models on participating teachers’ vicarious experiences, thereby enhancing their teaching self-efficacy. Finally, the design of activity contexts fully leverages the dual advantages of online and offline activities, meeting teachers’ needs for real-time face-to-face communication, on-site observation and experience, and hands-on operations, while also expanding the temporal and spatial context of professional development, making it more conducive to integrating with teachers’ daily classroom practices.
References:
[1] BANDURA A. Self-efficacy: Toward a Unifying Theory of Behavioral Change[J]. Advances in Behaviour Research & Therapy, 1977, 1(4):139-161.
[2] RIGGS I M, ENOCHS G. Toward the Development of an Elementary Teacher’s Science Teaching Efficacy Belief Instrument[J]. Science Education, 2010, 625-637.
[3] TSCHANNEN-MORAN M, HOY A W. The Differential Antecedents of Self-Efficacy Beliefs of Novice and Experienced Teachers[J]. Teaching and Teacher Education, 2007, 23(6): 944-956.
[4] HENSON, ROBIN K. From Adolescent Angst to Adulthood: Substantive Implications and Measurement Dilemmas in the Development of Teacher Efficacy Research[J]. Educational Psychologist, 2002, 37(3):137-150.
[5] KLASSEN R.M, TZE V M C, BETTS S M, et al. Teacher Efficacy Research 1998-2009: Signs of Progress or Unfulfilled Promise? [J]. Educational Psychology Review, 2011, 23(1):21-43.
[6] USHER E L, PAJARES F. Sources of Self-efficacy in School: Critical Review of the Literature and Future Directions[J]. Review of Educational Research, 2008, 78(4): 751-796.
[7] Lรผ Li. The Role of Vicarious Experience in Science Learning Self-efficacy[J]. Primary Science (Teacher Edition), 2020(9):60.
[8] MENEKSE M. Computer Science Teacher Professional Development in the United States: A Review of Studies Published Between 2004 and 2014[J]. Computer Science Education, 2015, 25(4):1-26.
[9] GUZDIAL M, ERICSON B, MCKLIN T, et al. Georgia Computes! An Intervention in a US State, with Formal and Informal Education in a Policy Context[J]. ACM Transactions on Computing Education, 2014, 14(2):1-29.
[10] FALKNER K, VIVIAN R, FALKNER N, et al. Reflecting on Three Offerings of a Community-Centric MOOC for K-6 Computer Science Teachers[R]. The 2017 ACM SIGCSE Technical Symposium, 2017.
[11] HADUONG P, BRENNAN K. Helping K-12 Teachers Get Unstuck with Scratch: The Design of an Online Professional Learning Experience[R]. The 50th ACM Technical Symposium, 2019.
[12] KOSMAS P. Online Sharing of Knowledge Among In-service Teachers for Professional Development Purposes[R]. International Conference on Education Technology & Computers. 2017๏ผ162-166.
[13] GOODE J, MARGOLIS J, CHAPMAN G. Curriculum is not enough: the educational theory and research foundation of the exploring computer science professional development model[R]. Proceedings of the 45th ACM technical symposium on Computer science education, 2014๏ผ493-498.
[14] MASHIKHI K M A, SOLIMAN S A. A Training Program in the Framework of Rogers’Theory of Blended Learning: A Case of Teachers of Information Technology in Oman [J]. Arab World English Journal, 2019, 9(4): 361-379.
[15] LAZARINIS F, KARACHRISTOS C V, STAVROPOULOS E C, et al. A Blended Learning Course for Playfully Teaching Programming Concepts to School Teachers[J]. Education and Information Technologies, 2019, 24(2), 1237-1249.
[16] KOSMAS P. Online Sharing of Knowledge Among In-service Teachers for Professional Development Purposes[C]. International Conference on Education Technology and Computers, 2017๏ผ162-166.
[17] State Council. Notice of the State Council on Printing and Distributing the New Generation Artificial Intelligence Development Plan[EB/OL]. 2017-07-26[2022-04-02]. http://www.gov.cn/zhengce/content/2017-07/20/content_5211996.
[18] Xie Zhongxin, Cao Yanglu, Li Ying. Research on the Design of AI Curriculum Content for K-12 Schools[J]. China Educational Technology, 2019, (4):17-22.
[19] Zhang Zhixin, Du Hui, Gao Lu, et al. Current Situation, Problems, and Countermeasures of AI Curriculum Construction in K-12 Schools in Developed Areas: A Case Study of a โNew First-tierโ City[J]. China Educational Technology, 2020๏ผ(9):40-49.
[20] EVANS R, LUFT J, CZERNIAK C, et al. The Role of Science Teachers’ Beliefs in International Classrooms: From Teacher Actions to Student Learning[M]. Sense Publishers, 2014.
[21] DILEKLI Y, TEZCI E. A Cross-cultural Study: Teachers’ Self-Efficacy Beliefs for Teaching Thinking Skills[J]. Thinking Skills and Creativity, 2019, 35:100624.
[22] TABER K S. BARRY J.FRASER, et al. McRobbie (eds): Second International Handbook of Science Education[J]. Science & Education, 2015, 24:319-33.
[23] Desimone L M. Improving Impact Studies of Teachers’ Professional Development: Toward Better Conceptualizations and Measures[J]. Educational Researcher, 2009, 38(3):181-199.
[24] Zhao Wei. Teaching Self-efficacy of Information Technology Teachers: Connotation, Status, and Countermeasures[J]. China Educational Technology, 2010(6):96-99.
[25] YADAV A, LISHINSKI A, SANDS P. Self-efficacy Profiles for Computer Science Teachers[C]. SIGCSE ’21: The 52nd ACM Technical Symposium on Computer Science Education, 2021.
[26] SCHWARZHAUPT R, LIU F, WILSON J, et al. Teachersโ Engagement and Self-Efficacy in a PK-12 Computer Science Teacher Virtual Community of Practice[J]. Journal of Computer Science Integration, 2021, 4(1): 1-14.
[27] ZHOU N, NGUYEN H, FISCHER C, et al. High School Teachers’ Self-efficacy in Teaching Computer Science[J]. ACM Transactions on Computing Education, 2020, 20(3):1-18.
[28] RICH P J, MASON S L, O’LEARY J. Measuring the Effect of Continuous Professional Development on Elementary Teachers’ Self-Efficacy to Teach Coding and Computational Thinking[J]. Computers & Education, 2021(3):104196.
[29] ZHAO M, ZHAO M, WANG X H, et al. Promoting Teaching Self-efficacy in Computational Thinking of School Teachers[C]. 2020 Ninth International Conference of Educational Innovation through Technology (EITT), 2020.
[30] JOCIUS R, JOSHI D, DONG Y, et al. Code, Connect, Create: The 3C Professional Development Model to Support Computational Thinking Infusion[C]. SIGCSE ’20: The 51st ACM Technical Symposium on Computer Science Education, 2020.
[31] MISHRA P, KOEHLER M J. Technological Pedagogical Content Knowledge (TPCK): Confronting the Wicked Problems of Teaching with Technology[J]. Proceedings of Society for Information Technology and Teacher Education International Conference, 2007:2214-2226.
[32] TOURETZKY D, GARDNER-MCCUNE C, MARTIN F, et al. Envisioning AI for K-12: What Should Every Child Know about AI? [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33: 9795-9799.
[33] DARLING-HAMMOND L, HYLER M E, GARDNER M. Effective teacher professional development[R]. Learning Policy Institute, 2017.
[34] GRAHAM J. Missing Data Analysis: Making It Work in the Real World[J]. Annual Review of Psychology, 2009, 60(1): 549-576.
[35] AKYUZ D. Measuring Technological Pedagogical Content Knowledge (TPACK) Through Performance Assessment[J]. Computers & Education, 2018, 125: 212-225.
[36] CHAI C S, KOH J H L. Changing Teachers’ TPACK and Design Beliefs Through the Scaffolded TPACK Lesson Design Model (STLDM)[J]. Learning: Research and Practice, 2017, 3(2):114-129.
[37] Yan Zhiming, Fu Jialiu, Zhu Youliang, et al. Integrating AI Technology into Subject Teaching Knowledge (AI-TPACK): Connotation, Teaching Practice, and Future Issues[J]. Journal of Distance Education, 2020, 38(5): 23-34.
[38] Deng Guomin, Li Yunchun, Zhu Yonghai. Reconstruction of Teacher Knowledge Structure Driven by “Artificial Intelligence + Education”: Discussing the AIPCEK Framework with Ethics and Its Development Model[J]. Journal of Distance Education, 2021, (1): 63-73.
[39] CHAI C S. Teacher Professional Development for Science, Technology, Engineering and Mathematics (STEM) Education: A Review from the Perspectives of Technological Pedagogical Content (TPACK) [J]. The Asia-Pacific Education Researcher, 2019, 28 (1): 5-13.
[40] BORKO H. Professional Development and Teacher Learning: Mapping the Terrain[J]. Educational Researcher, 2004, 33(8): 3-15.
[41] ALEATA H. Pedagogical Content Knowledge in Computing Education: A Review of the Research Literature[J]. Computer Science Education, 2018, 28:1-19.
[42] MING-YUEH H, JON-CHAO H, YUNG-WEI H. The Value of CK, PK, and PCK in Professional Development Programs Predicted by the Progressive Beliefs of Elementary School Teachers[J]. European Journal of Teacher Education, 2018, 41(4),448-462.