research on teachers professional development

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 08 September 2024

Longitudinal analysis of teacher self-efficacy evolution during a STEAM professional development program: a qualitative case study

• Haozhe Jiang   ORCID: orcid.org/0000-0002-7870-0993 1 ,
• Ritesh Chugh   ORCID: orcid.org/0000-0003-0061-7206 2 ,
• Xuesong Zhai   ORCID: orcid.org/0000-0002-4179-7859 1 , 3   nAff7 ,
• Ke Wang 4 &
• Xiaoqin Wang 5 , 6  

Humanities and Social Sciences Communications volume  11 , Article number:  1162 ( 2024 ) Cite this article

28 Accesses

Metrics details

Despite the widespread advocacy for the integration of arts and humanities (A&H) into science, technology, engineering, and mathematics (STEM) education on an international scale, teachers face numerous obstacles in practically integrating A&H into STEM teaching (IAT). To tackle the challenges, a comprehensive five-stage framework for teacher professional development programs focussed on IAT has been developed. Through the use of a qualitative case study approach, this study outlines the shifts in a participant teacher’s self-efficacy following their exposure to each stage of the framework. The data obtained from interviews and reflective analyses were analyzed using a seven-stage inductive method. The findings have substantiated the significant impact of a teacher professional development program based on the framework on teacher self-efficacy, evident in both individual performance and student outcomes observed over eighteen months. The evolution of teacher self-efficacy in IAT should be regarded as an open and multi-level system, characterized by interactions with teacher knowledge, skills and other entrenched beliefs. Building on our research findings, an enhanced model of teacher professional learning is proposed. The revised model illustrates that professional learning for STEAM teachers should be conceived as a continuous and sustainable process, characterized by the dynamic interaction among teaching performance, teacher knowledge, and teacher beliefs. The updated model further confirms the inseparable link between teacher learning and student learning within STEAM education. This study contributes to the existing body of literature on teacher self-efficacy, teacher professional learning models and the design of IAT teacher professional development programs.

Similar content being viewed by others

Primary and secondary school teachers’ perceptions of their social science training needs

Investigating how subject teachers transition to integrated STEM education: A hybrid qualitative study on primary and middle school teachers

The mediating role of teaching enthusiasm in the relationship between mindfulness, growth mindset, and psychological well-being of Chinese EFL teachers

Introduction.

In the past decade, there has been a surge in the advancement and widespread adoption of Science, Technology, Engineering, and Mathematics (STEM) education on a global scale (Jiang et al. 2021 ; Jiang et al. 2022 ; Jiang et al. 2023 ; Jiang et al. 2024a , b ; Zhan et al. 2023 ; Zhan and Niu 2023 ; Zhong et al. 2022 ; Zhong et al. 2024 ). Concurrently, there has been a growing chorus of advocates urging the integration of Arts and Humanities (A&H) into STEM education (e.g., Alkhabra et al. 2023 ; Land 2020 ; Park and Cho 2022 ; Uştu et al. 2021 ; Vaziri and Bradburn 2021 ). STEM education is frequently characterized by its emphasis on logic and analysis; however, it may be perceived as deficient in emotional and intuitive elements (Ozkan and Umdu Topsakal 2021 ). Through the integration of Arts and Humanities (A&H), the resulting STEAM approach has the potential to become more holistic, incorporating both rationality and emotional intelligence (Ozkan and Umdu Topsakal 2021 ). Many studies have confirmed that A&H can help students increase interest and develop their understanding of the contents in STEM fields, and thus, A&H can attract potential underrepresented STEM learners such as female students and minorities (Land 2020 ; Park and Cho 2022 ; Perignat and Katz-Buonincontro 2019 ). Despite the increasing interest in STEAM, the approaches to integrating A&H, which represent fundamentally different disciplines, into STEM are theoretically and practically ambiguous (Jacques et al. 2020 ; Uştu et al. 2021 ). Moreover, studies have indicated that the implementation of STEAM poses significant challenges, with STEM educators encountering difficulties in integrating A&H into their teaching practices (e.g., Boice et al. 2021 ; Duong et al. 2024 ; Herro et al. 2019 ; Jacques et al. 2020 ; Park and Cho 2022 ; Perignat and Katz-Buonincontro 2019 ). Hence, there is a pressing need to provide STEAM teachers with effective professional training.

Motivated by this gap, this study proposes a novel five-stage framework tailored for teacher professional development programs specifically designed to facilitate the integration of A&H into STEM teaching (IAT). Following the establishment of this framework, a series of teacher professional development programs were implemented. To explain the framework, a qualitative case study is employed, focusing on examining a specific teacher professional development program’s impact on a pre-service teacher’s self-efficacy. The case narratives, with a particular focus on the pre-service teacher’s changes in teacher self-efficacy, are organized chronologically, delineating stages before and after each stage of the teacher professional development program. More specifically, meaningful vignettes of the pre-service teacher’s learning and teaching experiences during the teacher professional development program are offered to help understand the five-stage framework. This study contributes to understanding teacher self-efficacy, teacher professional learning model and the design of IAT teacher professional development programs.

Theoretical background

The conceptualization of steam education.

STEM education can be interpreted through various lenses (e.g., Jiang et al. 2021 ; English 2016 ). As Li et al. (2020) claimed, on the one hand, STEM education can be defined as individual STEM disciplinary-based education (i.e., science education, technology education, engineering education and mathematics education). On the other hand, STEM education can also be defined as interdisciplinary or cross-disciplinary education where individual STEM disciplines are integrated (Jiang et al. 2021 ; English 2016 ). In this study, we view it as individual disciplinary-based education separately in science, technology, engineering and mathematics (English 2016 ).

STEAM education emerged as a new pedagogy during the Americans for the Arts-National Policy Roundtable discussion in 2007 (Perignat and Katz-Buonincontro 2019 ). This pedagogy was born out of the necessity to enhance students’ engagement, foster creativity, stimulate innovation, improve problem-solving abilities, and cultivate employability skills such as teamwork, communication and adaptability (Perignat and Katz-Buonincontro 2019 ). In particular, within the framework of STEAM education, the ‘A’ should be viewed as a broad concept that represents arts and humanities (A&H) (Herro and Quigley 2016 ; de la Garza 2021 , Park and Cho 2022 ). This conceptualization emphasizes the need to include humanities subjects alongside arts (Herro and Quigley 2016 ; de la Garza 2021 ; Park and Cho 2022 ). Sanz-Camarero et al. ( 2023 ) listed some important fields of A&H, including physical arts, fine arts, manual arts, sociology, politics, philosophy, history, psychology and so on.

In general, STEM education does not necessarily entail the inclusion of all STEM disciplines collectively (Ozkan and Umdu Topsakal 2021 ), and this principle also applies to STEAM education (Gates 2017 ; Perignat and Katz-Buonincontro 2019 ; Quigley et al. 2017 ; Smith and Paré 2016 ). As an illustration, Smith and Paré ( 2016 ) described a STEAM activity in which pottery (representing A&H) and mathematics were integrated, while other STEAM elements such as science, technology and engineering were not included. In our study, STEAM education is conceptualized as an interdisciplinary approach that involves the integration of one or more components of A&H into one or more STEM school subjects within educational activities (Ozkan and Umdu Topsakal 2021 ; Vaziri and Bradburn 2021 ). Notably, interdisciplinary collaboration entails integrating one or more elements from arts and humanities (A&H) with one or more STEM school subjects, cohesively united by a shared theme while maintaining their distinct identities (Perignat and Katz-Buonincontro 2019 ).

In our teacher professional development programs, we help mathematics, technology, and science pre-service teachers integrate one component of A&H into their disciplinary-based teaching practices. For instance, we help mathematics teachers integrate history (a component of A&H) into mathematics teaching. In other words, in our study, integrating A&H into STEM teaching (IAT) can be defined as integrating one component of A&H into the teaching of one of the STEM school subjects. The components of A&H and the STEM school subject are brought together under a common theme, but each of them remains discrete. Engineering is not taught as an individual subject in the K-12 curriculum in mainland China. Therefore, A&H is not integrated into engineering teaching in our teacher professional development programs.

Self-efficacy and teacher self-efficacy

Self-efficacy was initially introduced by Bandura ( 1977 ) as a key concept within his social cognitive theory. Bandura ( 1997 ) defined self-efficacy as “people’s beliefs about their capabilities to produce designated levels of performance that exercise influence over events that affect their lives” (p. 71). Based on Bandura’s ( 1977 ) theory, Tschannen-Moran et al. ( 1998 ) defined the concept of teacher self-efficacy Footnote 1 as “a teacher’s belief in her or his ability to organize and execute the courses of action required to successfully accomplish a specific teaching task in a particular context” (p. 233). Blonder et al. ( 2014 ) pointed out that this definition implicitly included teachers’ judgment of their ability to bring about desired outcomes in terms of students’ engagement and learning. Moreover, OECD ( 2018 ) defined teacher self-efficacy as “the beliefs that teachers have of their ability to enact certain teaching behavior that influences students’ educational outcomes, such as achievement, interest, and motivation” (p. 51). This definition explicitly included two dimensions: teachers’ judgment of the ability related to their teaching performance (i.e., enacting certain teaching behavior) and their influence on student outcomes.

It is argued that teacher self-efficacy should not be regarded as a general or overarching construct (Zee et al. 2017 ; Zee and Koomen 2016 ). Particularly, in the performance-driven context of China, teachers always connect their beliefs in their professional capabilities with the educational outcomes of their students (Liu et al. 2018 ). Therefore, we operationally conceptualize teacher self-efficacy as having two dimensions: self-efficacy in individual performance and student outcomes (see Table 1 ).

Most importantly, given its consistent association with actual teaching performance and student outcomes (Bray-Clark and Bates 2003 ; Kelley et al. 2020 ), teacher self-efficacy is widely regarded as a pivotal indicator of teacher success (Kelley et al. 2020 ). Moreover, the enhancement of teaching self-efficacy reflects the effectiveness of teacher professional development programs (Bray-Clark and Bates 2003 ; Kelley et al. 2020 ; Wong et al. 2022 ; Zhou et al. 2023 ). For instance, Zhou et al. ( 2023 ) claimed that in STEM teacher education, effective teacher professional development programs should bolster teachers’ self-efficacy “in teaching the content in the STEM discipline” (p. 2).

It has been documented that teachers frequently experience diminished confidence and comfort when teaching subject areas beyond their expertise (Kelley et al. 2020 ; Stohlmann et al. 2012 ). This diminished confidence extends to their self-efficacy in implementing interdisciplinary teaching approaches, such as integrated STEM teaching and IAT (Kelley et al. 2020 ). For instance, Geng et al. ( 2019 ) found that STEM teachers in Hong Kong exhibited low levels of self-efficacy, with only 5.53% of teachers rating their overall self-efficacy in implementing STEM education as higher than a score of 4 out of 5. Additionally, Hunter-Doniger and Sydow ( 2016 ) found that teachers may experience apprehension and lack confidence when incorporating A&H elements into the classroom context, particularly within the framework of IAT. Considering the critical importance of teacher self-efficacy in STEM and STEAM education (Kelley et al. 2020 ; Zakariya, 2020 ; Zhou et al. 2023 ), it is necessary to explore effective measures, frameworks and teacher professional development programs to help teachers improve their self-efficacy regarding interdisciplinary teaching (e.g., IAT).

Teacher professional learning models

The relationship between teachers’ professional learning and students’ outcomes (such as achievements, skills and attitudes) has been a subject of extensive discussion and research for many years (Clarke and Hollingsworth 2002 ). For instance, Clarke and Hollingsworth ( 2002 ) proposed and validated the Interconnected Model of Professional Growth, which illustrates that teacher professional development is influenced by the interaction among four interconnected domains: the personal domain (teacher knowledge, beliefs and attitudes), the domain of practice (professional experimentation), the domain of consequence (salient outcomes), and the external domain (sources of information, stimulus or support). Sancar et al. ( 2021 ) emphasized that teachers’ professional learning or development never occurs independently. In practice, this process is inherently intertwined with many variables, including student outcomes, in various ways (Sancar et al. 2021 ). However, many current teacher professional development programs exclude real in-class teaching and fail to establish a comprehensive link between teachers’ professional learning and student outcomes (Cai et al. 2020 ; Sancar et al. 2021 ). Sancar et al. ( 2021 ) claimed that exploring the complex relationships between teachers’ professional learning and student outcomes should be grounded in monitoring and evaluating real in-class teaching, rather than relying on teachers’ self-assessment. It is essential to understand these relationships from a holistic perspective within the context of real classroom teaching (Sancar et al. 2021 ). However, as Sancar et al. ( 2021 ) pointed out, such efforts in teacher education are often considered inadequate. Furthermore, in the field of STEAM education, such efforts are further exacerbated.

Cai et al. ( 2020 ) proposed a teacher professional learning model where student outcomes are emphasized. This model was developed based on Cai ( 2017 ), Philipp ( 2007 ) and Thompson ( 1992 ). It has also been used and justified in a series of teacher professional development programs (e.g., Calabrese et al. 2024 ; Hwang et al. 2024 ; Marco and Palatnik 2024 ; Örnek and Soylu 2021 ). The model posits that teachers typically increase their knowledge and modify their beliefs through professional teacher learning, subsequently improving their classroom instruction, enhancing teaching performance, and ultimately fostering improved student learning outcomes (Cai et al. 2020 ). Notably, this model can be updated in several aspects. Firstly, prior studies have exhibited the interplay between teacher knowledge and beliefs (e.g., Basckin et al. 2021 ; Taimalu and Luik 2019 ). This indicates that the increase in teacher knowledge and the change in teacher belief may not be parallel. The two processes can be intertwined. Secondly, the Interconnected Model of Professional Growth highlights that the personal domain and the domain of practice are interconnected (Clarke and Hollingsworth 2002 ). Liu et al. ( 2022 ) also confirmed that improvements in classroom instruction may, in turn, influence teacher beliefs. This necessitates a reconsideration of the relationships between classroom instruction, teacher knowledge and teacher beliefs in Cai et al.’s ( 2020 ) model. Thirdly, the Interconnected Model of Professional Growth also exhibits the connections between the domain of consequence and the personal domain (Clarke and Hollingsworth 2002 ). Hence, the improvement of learning outcomes may signify the end of teacher learning. For instance, students’ learning feedback may be a vital source of teacher self-efficacy (Bandura 1977 ). Therefore, the improvement of student outcomes may, in turn, affect teacher beliefs. The aforementioned arguments highlight the need for an updated model that integrates Cai et al.’s ( 2020 ) teacher professional learning model with Clarke and Hollingsworth’s ( 2002 ) Interconnected Model of Professional Growth. This integration may provide a holistic view of the teacher’s professional learning process, especially within the complex contexts of STEAM teacher education.

The framework for teacher professional development programs of integrating arts and humanities into STEM teaching

In this section, we present a framework for IAT teacher professional development programs, aiming to address the practical challenges associated with STEAM teaching implementation. Our framework incorporates the five features of effective teacher professional development programs outlined by Archibald et al. ( 2011 ), Cai et al. ( 2020 ), Darling-Hammond et al. ( 2017 ), Desimone and Garet ( 2015 ) and Roth et al. ( 2017 ). These features include: (a) alignment with shared goals (e.g., school, district, and national policies and practice), (b) emphasis on core content and modeling of teaching strategies for the content, (c) collaboration among teachers within a community, (d) adequate opportunities for active learning of new teaching strategies, and (e) embedded follow-up and continuous feedback. It is worth noting that two concepts, namely community of practice and lesson study, have been incorporated into our framework. Below, we delineate how these features are reflected in our framework.

(a) The Chinese government has issued a series of policies to facilitate STEAM education in K-12 schools (Jiang et al. 2021 ; Li and Chiang 2019 ; Lyu et al. 2024 ; Ro et al. 2022 ). The new curriculum standards released in 2022 mandate that all K-12 teachers implement interdisciplinary teaching, including STEAM education. Our framework for teacher professional development programs, which aims to help teachers integrate A&H into STEM teaching, closely aligns with these national policies and practices supporting STEAM education in K-12 schools.

(b) The core content of the framework is IAT. Specifically, as A&H is a broad concept, we divide it into several subcomponents, such as history, culture, and visual and performing arts (e.g., drama). We are implementing a series of teacher professional development programs to help mathematics, technology and science pre-service teachers integrate these subcomponents of A&H into their teaching Footnote 2 . Notably, pre-service teachers often lack teaching experience, making it challenging to master and implement new teaching strategies. Therefore, our framework provides five step-by-step stages designed to help them effectively model the teaching strategies of IAT.

(c) Our framework advocates for collaboration among teachers within a community of practice. Specifically, a community of practice is “a group of people who share an interest in a domain of human endeavor and engage in a process of collective learning that creates bonds between them” (Wenger et al. 2002 , p. 1). A teacher community of practice can be considered a group of teachers “sharing and critically observing their practices in growth-promoting ways” (Näykki et al. 2021 , p. 497). Long et al. ( 2021 ) claimed that in a teacher community of practice, members collaboratively share their teaching experiences and work together to address teaching problems. Our community of practice includes three types of members. (1) Mentors: These are professors and experts with rich experience in helping pre-service teachers practice IAT. (2) Pre-service teachers: Few have teaching experience before the teacher professional development programs. (3) In-service teachers: All in-service teachers are senior teachers with rich teaching experience. All the members work closely together to share and improve their IAT practice. Moreover, our community includes not only mentors and in-service teachers but also pre-service teachers. We encourage pre-service teachers to collaborate with experienced in-service teachers in various ways, such as developing IAT lesson plans, writing IAT case reports and so on. In-service teachers can provide cognitive and emotional support and share their practical knowledge and experience, which may significantly benefit the professional growth of pre-service teachers (Alwafi et al. 2020 ).

(d) Our framework offers pre-service teachers various opportunities to engage in lesson study, allowing them to actively design and implement IAT lessons. Based on the key points of effective lesson study outlined by Akiba et al. ( 2019 ), Ding et al. ( 2024 ), and Takahashi and McDougal ( 2016 ), our lesson study incorporates the following seven features. (1) Study of IAT materials: Pre-service teachers are required to study relevant IAT materials under the guidance of mentors. (2) Collaboration on lesson proposals: Pre-service teachers should collaborate with in-service teachers to develop comprehensive lesson proposals. (3) Observation and data collection: During the lesson, pre-service teachers are required to carefully observe and collect data on student learning and development. (4) Reflection and analysis: Pre-service teachers use the collected data to reflect on the lesson and their teaching effects. (5) Lesson revision and reteaching: If needed, pre-service teachers revise and reteach the lesson based on their reflections and data analysis. (6) Mentor and experienced in-service teacher involvement: Mentors and experienced in-service teachers, as knowledgeable others, are involved throughout the lesson study process. (7) Collaboration on reporting: Pre-service teachers collaborate with in-service teachers to draft reports and disseminate the results of the lesson study. Specifically, recognizing that pre-service teachers often lack teaching experience, we do not require them to complete all the steps of lesson study independently at once. Instead, we guide them through the lesson study process in a step-by-step manner, allowing them to gradually build their IAT skills and confidence. For instance, in Stage 1, pre-service teachers primarily focus on studying IAT materials. In Stage 2, they develop lesson proposals, observe and collect data, and draft reports. However, the implementation of IAT lessons is carried out by in-service teachers. This approach prevents pre-service teachers from experiencing failures due to their lack of teaching experience. In Stage 3, pre-service teachers implement, revise, and reteach IAT lessons, experiencing the lesson study process within a simulated environment. In Stage 4, pre-service teachers engage in lesson study in an actual classroom environment. However, their focus is limited to one micro-course during each lesson study session. It is not until the fifth stage that they experience a complete lesson study in an actual classroom environment.

(e) Our teacher professional development programs incorporate assessments specifically designed to evaluate pre-service teachers’ IAT practices. We use formative assessments to measure their understanding and application of IAT strategies. Pre-service teachers receive ongoing and timely feedback from peers, mentors, in-service teachers, and students, which helps them continuously refine their IAT practices throughout the program. Recognizing that pre-service teachers often have limited contact with real students and may not fully understand students’ learning needs, processes and outcomes, our framework requires them to actively collect and analyze student feedback. By doing so, they can make informed improvements to their instructional practice based on student feedback.

After undergoing three rounds of theoretical and practical testing and revision over the past five years, we have successfully finalized the optimization of the framework design (Zhou 2021 ). Throughout each cycle, we collected feedback from both participants and researchers on at least three occasions. Subsequently, we analyzed this feedback and iteratively refined the framework. For example, we enlisted the participation of in-service teachers to enhance the implementation of STEAM teaching, extended practice time through micro-teaching sessions, and introduced a stage of micro-course development within the framework to provide more opportunities for pre-service teachers to engage with real teaching situations. In this process, we continuously improved the coherence between each stage of the framework, ensuring that they mutually complement one another. The five-stage framework is described as follows.

Stage 1 Literature study

Pre-service teachers are provided with a series of reading materials from A&H. On a weekly basis, two pre-service teachers are assigned to present their readings and reflections to the entire group, followed by critical discussions thereafter. Mentors and all pre-service teachers discuss and explore strategies for translating the original A&H materials into viable instructional resources suitable for classroom use. Subsequently, pre-service teachers select topics of personal interest for further study under mentor guidance.

Stage 2 Case learning

Given that pre-service teachers have no teaching experience, collaborative efforts between in-service teachers and pre-service teachers are undertaken to design IAT lesson plans. Subsequently, the in-service teachers implement these plans. Throughout this process, pre-service teachers are afforded opportunities to engage in lesson plan implementation. Figure 1 illustrates the role of pre-service teachers in case learning. In the first step, pre-service teachers read about materials related to A&H, select suitable materials, and report their ideas on IAT lesson design to mentors, in-service teachers, and fellow pre-service teachers.

Note: A&H refers to arts and humanities.

In the second step, they liaise with the in-service teachers responsible for implementing the lesson plan, discussing the integration of A&H into teaching practices. Pre-service teachers then analyze student learning objectives aligned with curriculum standards, collaboratively designing the IAT lesson plan with in-service teachers. Subsequently, pre-service teachers present lesson plans for feedback from mentors and other in-service teachers.

In the third step, pre-service teachers observe the lesson plan’s implementation, gathering and analyzing feedback from students and in-service teachers using an inductive approach (Merriam 1998 ). Feedback includes opinions on the roles and values of A&H, perceptions of the teaching effect, and recommendations for lesson plan implementation and modification. The second and third steps may iterate multiple times to refine the IAT lesson plan. In the fourth step, pre-service teachers consolidate all data, including various versions of teaching instructions, classroom videos, feedback, and discussion notes, composing reflection notes. Finally, pre-service teachers collaborate with in-service teachers to compile the IAT case report and submit it for publication.

Stage 3 Micro-teaching

Figure 2 illustrates the role of pre-service teachers in micro-teaching. Before entering the micro-classrooms Footnote 3 , all the discussions and communications occur within the pre-service teacher group, excluding mentors and in-service teachers. After designing the IAT lesson plan, pre-service teachers take turns implementing 40-min lesson plans in a simulated micro-classroom setting. Within this simulated environment, one pre-service teacher acts as the teacher, while others, including mentors, in-service teachers, and other fellow pre-service teachers, assume the role of students Footnote 4 . Following the simulated teaching, the implementer reviews the video of their session and self-assesses their performance. Subsequently, the implementer receives feedback from other pre-service teachers, mentors, and in-service teachers. Based on this feedback, the implementer revisits steps 2 and 3, revising the lesson plan and conducting the simulated teaching again. This iterative process typically repeats at least three times until the mentors, in-service teachers, and other pre-service teachers are satisfied with the implementation of the revised lesson plan. Finally, pre-service teachers complete reflection notes and submit a summary of their reflections on the micro-teaching experience. Each pre-service teacher is required to choose at least three topics and undergo at least nine simulated teaching sessions.

Stage 4 Micro-course development

While pre-service teachers may not have the opportunity to execute the whole lesson plans in real classrooms, they can design and create five-minute micro-courses Footnote 5 before class, subsequently presenting these videos to actual students. The process of developing micro-courses closely mirrors that of developing IAT cases in the case learning stage (see Fig. 1 ). However, in Step 3, pre-service teachers assume dual roles, not only as observers of IAT lesson implementation but also as implementers of a five-minute IAT micro-course.

Stage 5 Classroom teaching

Pre-service teachers undertake the implementation of IAT lesson plans independently, a process resembling micro-teaching (see Fig. 2 ). However, pre-service teachers engage with real school students in partner schools Footnote 6 instead of simulated classrooms. Furthermore, they collect feedback not only from the mentors, in-service teachers, and fellow pre-service teachers but also from real students.

To provide our readers with a better understanding of the framework, we provide meaningful vignettes of a pre-service teacher’s learning and teaching experiences in one of the teacher professional development programs based on the framework. In addition, we choose teacher self-efficacy as an indicator to assess the framework’s effectiveness, detailing the pre-service teacher’s changes in teacher self-efficacy.

Research design

Research method.

Teacher self-efficacy can be measured both quantitatively and qualitatively (Bandura 1986 , 1997 ; Lee and Bobko 1994 ; Soprano and Yang 2013 ; Unfried et al. 2022 ). However, researchers and theorists in the area of teacher self-efficacy have called for more qualitative and longitudinal studies (Klassen et al. 2011 ). As some critiques stated, most studies were based on correlational and cross-sectional data obtained from self-report surveys, and qualitative studies of teacher efficacy were overwhelmingly neglected (Henson 2002 ; Klassen et al. 2011 ; Tschannen-Moran et al. 1998 ; Xenofontos and Andrews 2020 ). There is an urgent need for more longitudinal studies to shed light on the development of teacher efficacy (Klassen et al. 2011 ; Xenofontos and Andrews 2020 ).

This study utilized a longitudinal qualitative case study methodology to delve deeply into the context (Jiang et al. 2021 ; Corden and Millar 2007 ; Dicks et al. 2023 ; Henderson et al. 2012 ; Matusovich et al. 2010 ; Shirani and Henwood 2011 ), presenting details grounded in real-life situations and analyzing the inner relationships rather than generalize findings about the change of a large group of pre-service teachers’ self-efficacy.

Participant

This study forms a component of a broader multi-case research initiative examining teachers’ professional learning in the STEAM teacher professional development programs in China (Jiang et al. 2021 ; Wang et al. 2018 ; Wang et al. 2024 ). Within this context, one participant, Shuitao (pseudonym), is selected and reported in this current study. Shuitao was a first-year graduate student at a first-tier Normal university in Shanghai, China. Normal universities specialize in teacher education. Her graduate major was mathematics curriculum and instruction. Teaching practice courses are offered to students in this major exclusively during their third year of study. The selection of Shuitao was driven by three primary factors. Firstly, Shuitao attended the entire teacher professional development program and actively engaged in nearly all associated activities. Table 2 illustrates the timeline of the five stages in which Shuitao was involved. Secondly, her undergraduate major was applied mathematics, which was not related to mathematics teaching Footnote 7 . She possessed no prior teaching experience and had not undergone any systematic study of IAT before her involvement in the teacher professional development program. Thirdly, her other master’s courses during her first two years of study focused on mathematics education theory and did not include IAT Footnote 8 . Additionally, she scarcely participated in any other teaching practice outside of the teacher professional development program. As a pre-service teacher, Shuitao harbored a keen interest in IAT. Furthermore, she discovered that she possessed fewer teaching skills compared to her peers who had majored in education during their undergraduate studies. Hence, she had a strong desire to enhance her teaching skills. Consequently, Shuitao decided to participate in our teacher professional development program.

Shuitao was grouped with three other first-year graduate students during the teacher professional development program. She actively collaborated with them at every stage of the program. For instance, they advised each other on their IAT lesson designs, observed each other’s IAT practice and offered constructive suggestions for improvement.

Research question

Shuitao was a mathematics pre-service teacher who participated in one of our teacher professional development programs, focusing on integrating history into mathematics teaching (IHT) Footnote 9 . Notably, this teacher professional development program was designed based on our five-stage framework for teacher professional development programs of IAT. To examine the impact of this teacher professional development program on Shuitao’s self-efficacy related to IHT, this case study addresses the following research question:

What changes in Shuitao’s self-efficacy in individual performance regarding integrating history into mathematics teaching (SE-IHT-IP) may occur through participation in the teacher professional development program?

What changes in Shuitao’s self-efficacy in student outcomes regarding integrating history into mathematics teaching (SE-IHT-SO) may occur through participation in the teacher professional development program?

Data collection and analysis

Before Shuitao joined the teacher professional development program, a one-hour preliminary interview was conducted to guide her in self-narrating her psychological and cognitive state of IHT.

During the teacher professional development program, follow-up unstructured interviews were conducted once a month with Shuitao. All discussions in the development of IHT cases were recorded, Shuitao’s teaching and micro-teaching were videotaped, and the reflection notes, journals, and summary reports written by Shuitao were collected.

After completing the teacher professional development program, Shuitao participated in a semi-structured three-hour interview. The objectives of this interview were twofold: to reassess her self-efficacy and to explore the relationship between her self-efficacy changes and each stage of the teacher professional development program.

Interview data, discussions, reflection notes, journals, summary reports and videos, and analysis records were archived and transcribed before, during, and after the teacher professional development program.

In this study, we primarily utilized data from seven interviews: one conducted before the teacher professional development program, five conducted after each stage of the program, and one conducted upon completion of the program. Additionally, we reviewed Shuitao’s five reflective notes, which were written after each stage, as well as her final summary report that encompassed the entire teacher professional development program.

Merriam’s ( 1998 ) approach to coding data and inductive approach to retrieving possible concepts and themes were employed using a seven-stage method. Considering theoretical underpinnings in qualitative research is common when interpreting data (Strauss and Corbin 1990 ). First, a list based on our conceptual framework of teacher self-efficacy (see Table 1 ) was developed. The list included two codes (i.e., SE-IHT-IP and SE-IHT-SO). Second, all data were sorted chronologically, read and reread to be better understood. Third, texts were coded into multi-colored highlighting and comment balloons. Fourth, the data for groups of meanings, themes, and behaviors were examined. How these groups were connected within the conceptual framework of teacher self-efficacy was confirmed. Fifth, after comparing, confirming, and modifying, the selective codes were extracted and mapped onto the two categories according to the conceptual framework of teacher self-efficacy. Accordingly, changes in SE-IHT-IP and SE-IHT-SO at the five stages of the teacher professional development program were identified, respectively, and then the preliminary findings came (Strauss and Corbin 1990 ). In reality, in Shuitao’s narratives, SE-IHT-IP and SE-IHT-SO were frequently intertwined. Through our coding process, we differentiated between SE-IHT-IP and SE-IHT-SO, enabling us to obtain a more distinct understanding of how these two aspects of teacher self-efficacy evolved over time. This helped us address the two research questions effectively.

Reliability and validity

Two researchers independently analyzed the data to establish inter-rater reliability. The inter-rater reliability was established as kappa = 0.959. Stake ( 1995 ) suggested that the most critical assertions in a study require the greatest effort toward confirmation. In this study, three methods served this purpose and helped ensure the validity of the findings. The first way to substantiate the statement about the changes in self-efficacy was by revisiting each transcript to confirm whether the participant explicitly acknowledged the changes (Yin 2003 ). Such a check was repeated in the analysis of this study. The second way to confirm patterns in the data was by examining whether Shuitao’s statements were replicated in separate interviews (Morris and Usher 2011 ). The third approach involved presenting the preliminary conclusions to Shuitao and affording her the opportunity to provide feedback on the data and conclusions. This step aimed to ascertain whether we accurately grasped the true intentions of her statements and whether our subjective interpretations inadvertently influenced our analysis of her statements. Additionally, data from diverse sources underwent analysis by at least two researchers, with all researchers reaching consensus on each finding.

As each stage of our teacher professional development programs spanned a minimum of three months, numerous documented statements regarding the enhancement of Shuitao’s self-efficacy regarding IHT were recorded. Notably, what we present here offers only a concise overview of findings derived from our qualitative analysis. The changes in Shuitao’s SE-IHT-IP and SE-IHT-SO are organized chronologically, delineating the period before and during the teacher professional development program.

Before the teacher professional development program: “I have no confidence in IHT”

Before the teacher professional development program, Shuitao frequently expressed her lack of confidence in IHT. On the one hand, Shuitao expressed considerable apprehension about her individual performance in IHT. “How can I design and implement IHT lesson plans? I do not know anything [about it]…” With a sense of doubt, confusion and anxiety, Shuitao voiced her lack of confidence in her ability to design and implement an IHT case that would meet the requirements of the curriculum standards. Regarding the reasons for her lack of confidence, Shuitao attributed it to her insufficient theoretical knowledge and practical experience in IHT:

I do not know the basic approaches to IHT that I could follow… it is very difficult for me to find suitable historical materials… I am very confused about how to organize [historical] materials logically around the teaching goals and contents… [Furthermore,] I am [a] novice, [and] I have no IHT experience.

On the other hand, Shuitao articulated very low confidence in the efficacy of her IHT on student outcomes:

I think my IHT will have a limited impact on student outcomes… I do not know any specific effects [of history] other than making students interested in mathematics… In fact, I always think it is difficult for [my] students to understand the history… If students cannot understand [the history], will they feel bored?

This statement suggests that Shuitao did not fully grasp the significance of IHT. In fact, she knew little about the educational significance of history for students, and she harbored no belief that her IHT approach could positively impact students. In sum, her SE-IHT-SO was very low.

After stage 1: “I can do well in the first step of IHT”

After Stage 1, Shuitao indicated a slight improvement in her confidence in IHT. She attributed this improvement to her acquisition of theoretical knowledge in IHT, the approaches for selecting history-related materials, and an understanding of the educational value of history.

One of Shuitao’s primary concerns about implementing IHT before the teacher professional development program was the challenge of sourcing suitable history-related materials. However, after Stage 1, Shuitao explicitly affirmed her capability in this aspect. She shared her experience of organizing history-related materials related to logarithms as an example.

Recognizing the significance of suitable history-related materials in effective IHT implementation, Shuitao acknowledged that conducting literature studies significantly contributed to enhancing her confidence in undertaking this initial step. Furthermore, she expressed increased confidence in designing IHT lesson plans by utilizing history-related materials aligned with teaching objectives derived from the curriculum standards. In other words, her SE-IHT-IP was enhanced. She said:

After experiencing multiple discussions, I gradually know more about what kinds of materials are essential and should be emphasized, what kinds of materials should be adapted, and what kinds of materials should be omitted in the classroom instructions… I have a little confidence to implement IHT that could meet the requirements [of the curriculum standards] since now I can complete the critical first step [of IHT] well…

However, despite the improvement in her confidence in IHT following Stage 1, Shuitao also expressed some concerns. She articulated uncertainty regarding her performance in the subsequent stages of the teacher professional development program. Consequently, her confidence in IHT experienced only a modest increase.

After stage 2: “I participate in the development of IHT cases, and my confidence is increased a little bit more”

Following Stage 2, Shuitao reported further increased confidence in IHT. She attributed this growth to two main factors. Firstly, she successfully developed several instructional designs for IHT through collaboration with in-service teachers. These collaborative experiences enabled her to gain a deeper understanding of IHT approaches and enhance her pedagogical content knowledge in this area, consequently bolstering her confidence in her ability to perform effectively. Secondly, Shuitao observed the tangible impact of IHT cases on students in real classroom settings, which reinforced her belief in the efficacy of IHT. These experiences instilled in her a greater sense of confidence in her capacity to positively influence her students through her implementation of IHT. Shuitao remarked that she gradually understood how to integrate suitable history-related materials into her instructional designs (e.g., employ a genetic approach Footnote 10 ), considering it as the second important step of IHT. She shared her experience of developing IHT instructional design on the concept of logarithms. After creating several iterations of IHT instructional designs, Shuitao emphasized that her confidence in SE-IHT-IP has strengthened. She expressed belief in her ability to apply these approaches to IHT, as well as the pedagogical content knowledge of IHT, acquired through practical experience, in her future teaching endeavors. The following is an excerpt from the interview:

I learned some effective knowledge, skills, techniques and approaches [to IHT]… By employing these approaches, I thought I could [and] I had the confidence to integrate the history into instructional designs very well… For instance, [inspired] by the genetic approach, we designed a series of questions and tasks based on the history of logarithms. The introduction of the new concept of logarithms became very natural, and it perfectly met the requirements of our curriculum standards, [which] asked students to understand the necessity of learning the concept of logarithms…

Shuitao actively observed the classroom teaching conducted by her cooperating in-service teacher. She helped her cooperating in-service teacher in collecting and analyzing students’ feedback. Subsequently, discussions ensued on how to improve the instructional designs based on this feedback. The refined IHT instructional designs were subsequently re-implemented by the in-service teacher. After three rounds of developing IHT cases, Shuitao became increasingly convinced of the significance and efficacy of integrating history into teaching practices, as evidenced by the following excerpt:

The impacts of IHT on students are visible… For instance, more than 93% of the students mentioned in the open-ended questionnaires that they became more interested in mathematics because of the [historical] story of Napier… For another example, according to the results of our surveys, more than 75% of the students stated that they knew log a ( M  +  N ) = log a M  × log a N was wrong because of history… I have a little bit more confidence in the effects of my IHT on students.

This excerpt highlights that Shuitao’s SE-IHT-SO was enhanced. She attributed this enhancement to her realization of the compelling nature of history and her belief in her ability to effectively leverage its power to positively influence her students’ cognitive and emotional development. This also underscores the importance of reinforcing pre-service teachers’ awareness of the significance of history. Nonetheless, Shuiato elucidated that she still retained concerns regarding the effectiveness of her IHT implementation. Her following statement shed light on why her self-efficacy only experienced a marginal increase in this stage:

Knowing how to do it successfully and doing it successfully in practice are two totally different things… I can develop IHT instructional designs well, but I have no idea whether I can implement them well and whether I can introduce the history professionally in practice… My cooperation in-service teacher has a long history of teaching mathematics and gains rich experience in educational practices… If I cannot acquire some required teaching skills and capabilities, I still cannot influence my students powerfully.

After stage 3: “Practice makes perfect, and my SE-IHT-IP is steadily enhanced after a hit”

After successfully developing IHT instructional designs, the next critical step was the implementation of these designs. Drawing from her observations of her cooperating in-service teachers’ IHT implementations and discussions with other pre-service teachers, Shuitao developed her own IHT lesson plans. In Stage 3, she conducted simulated teaching sessions and evaluated her teaching performance ten times Footnote 11 . Shuitao claimed that her SE-IHT-IP steadily improved over the course of these sessions. According to Shuitao, two main processes in Stage 3 facilitated this steady enhancement of SE-IHT-IP.

On the one hand, through the repeated implementation of simulated teaching sessions, Shuitao’s teaching proficiency and fluency markedly improved. Shuitao first described the importance of teaching proficiency and fluency:

Since the detailed history is not included in our curriculum standards and textbooks, if I use my historical materials in class, I have to teach more contents than traditional teachers. Therefore, I have to teach proficiently so that teaching pace becomes a little faster than usual… I have to teach fluently so as to use each minute efficiently in my class. Otherwise, I cannot complete the teaching tasks required [by curriculum standards].

As Shuitao said, at the beginning of Stage 3, her self-efficacy even decreased because she lacked teaching proficiency and fluency and was unable to complete the required teaching tasks:

In the first few times of simulated teaching, I always needed to think for a second about what I should say next when I finish one sentence. I also felt very nervous when I stood in the front of the classrooms. This made my narration of the historical story between Briggs and Napier not fluent at all. I paused many times to look for some hints on my notes… All these made me unable to complete the required teaching tasks… My [teaching] confidence took a hit.

Shuitao quoted the proverb, “practice makes perfect”, and she emphasized that it was repeated practice that improved her teaching proficiency and fluency:

I thought I had no other choice but to practice IHT repeatedly… [At the end of Stage 3,] I could naturally remember most words that I should say when teaching the topics that I selected… My teaching proficiency and fluency was improved through my repeated review of my instructional designs and implementation of IHT in the micro-classrooms… With the improvement [of my teaching proficiency and fluency], I could complete the teaching tasks, and my confidence was increased as well.

In addition, Shuitao also mentioned that through this kind of self-exploration in simulated teaching practice, her teaching skills and capabilities (e.g., blackboard writing, abilities of language organization abilities, etc.) improved. This process was of great help to her enhancement of SE-IHT-IP.

On the other hand, Shuitao’s simulated teaching underwent assessment by herself, with mentors, in-service teachers and fellow pre-service teachers. This comprehensive evaluation process played a pivotal role in enhancing her individual performance and self-efficacy. Reflecting on this aspect, Shuitao articulated the following sentiments in one of her reflection reports:

By watching the videos, conducting self-assessment, and collecting feedback from others, I can understand what I should improve or emphasize in my teaching. [Then,] I think my IHT can better meet the requirements [of curriculum standards]… I think my teaching performance is getting better and better.

After stage 4: “My micro-courses influenced students positively, and my SE-IHT-SO is steadily enhanced”

In Stage 4, Shuitao commenced by creating 5-min micro-course videos. Subsequently, she played these videos in her cooperating in-service teachers’ authentic classroom settings and collected student feedback. This micro-course was played at the end of her cooperating in-service teachers’ lesson Footnote 12 . Shuitao wrote in her reflections that this micro-course of logarithms helped students better understand the nature of mathematics:

According to the results of our surveys, many students stated that they knew the development and evolution of the concept of logarithms is a long process and many mathematicians from different countries have contributed to the development of the concept of logarithms… This indicated that my micro-course helped students better understand the nature of mathematics… My micro-course about the history informed students that mathematics is an evolving and human subject and helped them understand the dynamic development of the [mathematics] concept…

Meanwhile, Shuitao’s micro-course positively influenced some students’ beliefs towards mathematics. As evident from the quote below, integrating historical context into mathematics teaching transformed students’ perception of the subject, boosting Shuitao’s confidence too.

Some students’ responses were very exciting… [O]ne [typical] response stated, he always regarded mathematics as abstract, boring, and dreadful subject; but after seeing the photos of mathematicians and great men and learning the development of the concept of logarithms through the micro-course, he found mathematics could be interesting. He wanted to learn more the interesting history… Students’ such changes made me confident.

Furthermore, during post-class interviews, several students expressed their recognition of the significance of the logarithms concept to Shuitao, attributing this realization to the insights provided by prominent figures in the micro-courses. They also conveyed their intention to exert greater effort in mastering the subject matter. This feedback made Shuitao believe that her IHT had the potential to positively influence students’ attitudes towards learning mathematics.

In summary, Stage 4 marked Shuitao’s first opportunity to directly impact students through her IHT in authentic classroom settings. Despite implementing only brief 5-min micro-courses integrating history during each session, the effectiveness of her short IHT implementation was validated by student feedback. Shuitao unequivocally expressed that students actively engaged with her micro-courses and that these sessions positively influenced them, including attitudes and motivation toward mathematics learning, understanding of mathematics concepts, and beliefs regarding mathematics. These collective factors contributed to a steady enhancement of her confidence in SE-IHT-SO.

After stage 5: “My overall self-efficacy is greatly enhanced”

Following Stage 5, Shuitao reported a significant increase in her overall confidence in IHT, attributing it to gaining mastery through successful implementations of IHT in real classroom settings. On the one hand, Shuitao successfully designed and executed her IHT lesson plans, consistently achieving the teaching objectives mandated by curriculum standards. This significantly enhanced her SE-IHT-IP. On the other hand, as Shuitao’s IHT implementation directly influenced her students, her confidence in SE-IHT-SO experienced considerable improvement.

According to Bandura ( 1997 ), mastery experience is the most powerful source of self-efficacy. Shuitao’s statements confirmed this. As she claimed, her enhanced SE-IHT-IP in Stage 5 mainly came from the experience of successful implementations of IHT in real classrooms:

[Before the teacher professional development program,] I had no idea about implementing IHT… Now, I successfully implemented IHT in senior high school [classrooms] many times… I can complete the teaching tasks and even better completed the teaching objectives required [by the curriculum standards]… The successful experience greatly enhances my confidence to perform well in my future implementation of IHT… Yeah, I think the successful teaching practice experience is the strongest booster of my confidence.

At the end of stage 5, Shuitao’s mentors and in-service teachers gave her a high evaluation. For instance, after Shuitao’s IHT implementation of the concept of logarithms, all mentors and in-service teachers consistently provided feedback that her IHT teaching illustrated the necessity of learning the concept of logarithms and met the requirements of the curriculum standards very well. This kind of verbal persuasion (Bandura 1997 ) enhanced her SE-IHT-IP.

Similarly, Shuitao’s successful experience of influencing students positively through IHT, as one kind of mastery experience, powerfully enhanced her SE-IHT-SO. She described her changes in SE-IHT-SO as follows:

I could not imagine my IHT could be so influential [before]… But now, my IHT implementation directly influenced students in so many aspects… When I witnessed students’ real changes in various cognitive and affective aspects, my confidence was greatly improved.

Shuitao described the influence of her IHT implementation of the concept of logarithms on her students. The depiction is grounded in the outcomes of surveys conducted by Shuitao following her implementation. Shuitao asserted that these results filled her with excitement and confidence regarding her future implementation of IHT.

In summary, following Stage 5 of the teacher professional development program, Shuitao experienced a notable enhancement in her overall self-efficacy, primarily attributed to her successful practical experience in authentic classroom settings during this stage.

A primary objective of our teacher professional development programs is to equip pre-service teachers with the skills and confidence needed to effectively implement IAT. Our findings show that one teacher professional development program, significantly augmented a participant’s TSE-IHT across two dimensions: individual performance and student outcomes. Considering the pressing need to provide STEAM teachers with effective professional training (e.g., Boice et al. 2021 ; Duong et al. 2024 ; Herro et al. 2019 ; Jacques et al. 2020 ; Park and Cho 2022 ; Perignat and Katz-Buonincontro 2019 ), the proposed five-stage framework holds significant promise in both theoretical and practical realms. Furthermore, this study offers a viable solution to address the prevalent issue of low levels of teacher self-efficacy in interdisciplinary teaching, including IAT, which is critical in STEAM education (Zhou et al. 2023 ). This study holds the potential to make unique contributions to the existing body of literature on teacher self-efficacy, teacher professional learning models and the design of teacher professional development programs of IAT.

Firstly, this study enhances our understanding of the development of teacher self-efficacy. Our findings further confirm the complexity of the development of teacher self-efficacy. On the one hand, the observed enhancement of the participant’s teacher self-efficacy did not occur swiftly but unfolded gradually through a protracted, incremental process. Moreover, it is noteworthy that the participant’s self-efficacy exhibited fluctuations, underscoring that the augmentation of teacher self-efficacy is neither straightforward nor linear. On the other hand, the study elucidated that the augmentation of teacher self-efficacy constitutes an intricate, multi-level system that interacts with teacher knowledge, skills, and other beliefs. This finding resonates with prior research on teacher self-efficacy (Morris et al. 2017 ; Xenofontos and Andrews 2020 ). For example, our study revealed that Shuitao’s enhancement of SE-IHT-SO may always be interwoven with her continuous comprehension of the significance of the A&H in classroom settings. Similarly, the participant progressively acknowledged the educational value of A&H in classroom contexts in tandem with the stepwise enhancement of SE-IHT-SO. Factors such as the participant’s pedagogical content knowledge of IHT, instructional design, and teaching skills were also identified as pivotal components of SE-IHT-IP. This finding corroborates Morris and Usher ( 2011 ) assertion that sustained improvements in self-efficacy stem from developing teachers’ skills and knowledge. With the bolstering of SE-IHT-IP, the participant’s related teaching skills and content knowledge also exhibited improvement.

Methodologically, many researchers advocate for qualitative investigations into self-efficacy (e.g., Philippou and Pantziara 2015; Klassen et al. 2011 ; Wyatt 2015 ; Xenofontos and Andrews 2020 ). While acknowledging limitations in sample scope and the generalizability of the findings, this study offers a longitudinal perspective on the stage-by-stage development of teacher self-efficacy and its interactions with different factors (i.e., teacher knowledge, skills, and beliefs), often ignored by quantitative studies. Considering that studies of self-efficacy have been predominantly quantitative, typically drawing on survey techniques and pre-determined scales (Xenofontos and Andrews, 2020 ; Zhou et al. 2023 ), this study highlights the need for greater attention to qualitative studies so that more cultural, situational and contextual factors in the development of self-efficacy can be captured.

Our study provides valuable practical implications for enhancing pre-service teachers’ self-efficacy. We conceptualize teacher self-efficacy in two primary dimensions: individual performance and student outcomes. On the one hand, pre-service teachers can enhance their teaching qualities, boosting their self-efficacy in individual performance. The adage “practice makes perfect” underscores the necessity of ample teaching practice opportunities for pre-service teachers who lack prior teaching experience. Engaging in consistent and reflective practice helps them develop confidence in their teaching qualities. On the other hand, pre-service teachers should focus on positive feedback from their students, reinforcing their self-efficacy in individual performance. Positive student feedback serves as an affirmation of their teaching effectiveness and encourages continuous improvement. Furthermore, our findings highlight the significance of mentors’ and peers’ positive feedback as critical sources of teacher self-efficacy. Mentors and peers play a pivotal role in the professional growth of pre-service teachers by actively encouraging them and recognizing their teaching achievements. Constructive feedback from experienced mentors and supportive peers fosters a collaborative learning environment and bolsters the self-confidence of pre-service teachers. Additionally, our research indicates that pre-service teachers’ self-efficacy may fluctuate. Therefore, mentors should be prepared to help pre-service teachers manage teaching challenges and setbacks, and alleviate any teaching-related anxiety. Mentors can help pre-service teachers build resilience and maintain a positive outlook on their teaching journey through emotional support and guidance. Moreover, a strong correlation exists between teacher self-efficacy and teacher knowledge and skills. Enhancing pre-service teachers’ knowledge base and instructional skills is crucial for bolstering their overall self-efficacy.

Secondly, this study also responds to the appeal to understand teachers’ professional learning from a holistic perspective and interrelate teachers’ professional learning process with student outcome variables (Sancar et al. 2021 ), and thus contributes to the understanding of the complexity of STEAM teachers’ professional learning. On the one hand, we have confirmed Cai et al.’s ( 2020 ) teacher professional learning model in a new context, namely STEAM teacher education. Throughout the teacher professional development program, the pre-service teacher, Shuitao, demonstrated an augmentation in her knowledge, encompassing both content knowledge and pedagogical understanding concerning IHT. Moreover, her beliefs regarding IHT transformed as a result of her engagement in teacher learning across the five stages. This facilitated her in executing effective IHT teaching and improving her students’ outcomes. On the other hand, notably, in our studies (including this current study and some follow-up studies), student feedback is a pivotal tool to assist teachers in discerning the impact they are effectuating. This enables pre-service teachers to grasp the actual efficacy of their teaching efforts and subsequently contributes significantly to the augmentation of their self-efficacy. Such steps have seldom been conducted in prior studies (e.g., Cai et al. 2020 ), where student outcomes are often perceived solely as the results of teachers’ instruction rather than sources informing teacher beliefs. Additionally, this study has validated both the interaction between teaching performance and teacher beliefs and between teacher knowledge and teacher beliefs. These aspects were overlooked in Cai et al.’s ( 2020 ) model. More importantly, while Clarke and Hollingsworth’s ( 2002 ) Interconnected Model of Professional Growth illustrates the connections between the domain of consequence and the personal domain, as well as between the personal domain and the domain of practice, it does not adequately clarify the complex relationships among the factors within the personal domain (e.g., the interaction between teacher knowledge and teacher beliefs). Therefore, our study also supplements Clarke and Hollingsworth’s ( 2002 ) model by addressing these intricacies. Based on our findings, an updated model of teacher professional learning has been proposed, as shown in Fig. 3 . This expanded model indicates that teacher learning should be an ongoing and sustainable process, with the enhancement of student learning not marking the conclusion of teacher learning, but rather serving as the catalyst for a new phase of learning. In this sense, we advocate for further research to investigate the tangible impacts of teacher professional development programs on students and how those impacts stimulate subsequent cycles of teacher learning.

Note: Paths in blue were proposed by Cai et al. ( 2020 ), and paths in yellow are proposed and verified in this study.

Thirdly, in light of the updated model of teacher professional learning (see Fig. 3 ), this study provides insights into the design of teacher professional development programs of IAT. According to Huang et al. ( 2022 ), to date, very few studies have set goals to “develop a comprehensive understanding of effective designs” for STEM (or STEAM) teacher professional development programs (p. 15). To fill this gap, this study proposes a novel and effective five-stage framework for teacher professional development programs of IAT. This framework provides a possible and feasible solution to the challenges of STEAM teacher professional development programs’ design and planning, and teachers’ IAT practice (Boice et al. 2021 ; Herro et al. 2019 ; Jacques et al. 2020 ; Park and Cho 2022 ; Perignat and Katz-Buonincontro 2019 ).

Specifically, our five-stage framework incorporates at least six important features. Firstly, teacher professional development programs should focus on specific STEAM content. Given the expansive nature of STEAM, teacher professional development programs cannot feasibly encompass all facets of its contents. Consistent with recommendations by Cai et al. ( 2020 ), Desimone et al. ( 2002 ) and Garet et al. ( 2001 ), an effective teacher professional development program should prioritize content focus. Our five-stage framework is centered on IAT. Throughout an 18-month duration, each pre-service teacher is limited to selecting one subcomponent of A&H, such as history, for integration into their subject teaching (i.e., mathematics teaching, technology teaching or science teaching) within one teacher professional development program. Secondly, in response to the appeals that teacher professional development programs should shift from emphasizing teaching and instruction to emphasizing student learning (Cai et al. 2020 ; Calabrese et al. 2024 ; Hwang et al. 2024 ; Marco and Palatnik 2024 ; Örnek and Soylu 2021 ), our framework requires pre-service teachers to pay close attention to the effects of IAT on student learning outcomes, and use students’ feedback as the basis of improving their instruction. Thirdly, prior studies found that teacher education with a preference for theory led to pre-service teachers’ dissatisfaction with the quality of teacher professional development program and hindered the development of pre-service teachers’ teaching skills and teaching beliefs, which also widened the gap between theory and practice (Hennissen et al. 2017 ; Ord and Nuttall 2016 ). In this regard, our five-stage framework connects theory and teaching practice closely. In particular, pre-service teachers can experience the values of IAT not only through theoretical learning but also through diverse teaching practices. Fourthly, we build a teacher community of practice tailored for pre-service teachers. Additionally, we aim to encourage greater participation of in-service teachers in such teacher professional development programs designed for pre-service educators in STEAM teacher education. By engaging in such programs, in-service teachers can offer valuable teaching opportunities for pre-service educators and contribute their insights and experiences from teaching practice. Importantly, pre-service teachers stand to gain from the in-service teachers’ familiarity with textbooks, subject matter expertise, and better understanding of student dynamics. Fifthly, our five-stage framework lasts for an extended period, spanning 18 months. This duration ensures that pre-service teachers engage in a sustained and comprehensive learning journey. Lastly, our framework facilitates a practical understanding of “integration” by offering detailed, sequential instructions for blending two disciplines in teaching. For example, our teacher professional development programs prioritize systematic learning of pedagogical theories and simulated teaching experiences before pre-service teachers embark on real STEAM teaching endeavors. This approach is designed to mitigate the risk of unsuccessful experiences during initial teaching efforts, thereby safeguarding pre-service teachers’ teacher self-efficacy. Considering the complexity of “integration” in interdisciplinary teaching practices, including IAT (Han et al. 2022 ; Ryu et al. 2019 ), we believe detailed stage-by-stage and step-by-step instructions are crucial components of relevant pre-service teacher professional development programs. Notably, this aspect, emphasizing structural instructional guidance, has not been explicitly addressed in prior research (e.g., Cai et al. 2020 ). Figure 4 illustrates the six important features outlined in this study, encompassing both established elements and the novel addition proposed herein, describing an effective teacher professional development program.

Note: STEAM refers to science, technology, engineering, arts and humanities, and mathematics.

The successful implementation of this framework is also related to the Chinese teacher education system and cultural background. For instance, the Chinese government has promoted many university-school collaboration initiatives, encouraging in-service teachers to provide guidance and practical opportunities for pre-service teachers (Lu et al. 2019 ). Influenced by Confucian values emphasizing altruism, many experienced in-service teachers in China are eager to assist pre-service teachers, helping them better realize their teaching career aspirations. It is reported that experienced in-service teachers in China show significantly higher motivation than their international peers when mentoring pre-service teachers (Lu et al. 2019 ). Therefore, for the successful implementation of this framework in other countries, it is crucial for universities to forge close collaborative relationships with K-12 schools and actively involve K-12 teachers in pre-service teacher education.

Notably, approximately 5% of our participants dropped out midway as they found that the IAT practice was too challenging or felt overwhelmed by the number of required tasks in the program. Consequently, we are exploring options to potentially simplify this framework in future iterations.

Without minimizing the limitations of this study, it is important to recognize that a qualitative longitudinal case study can be a useful means of shedding light on the development of a pre-service STEAM teacher’s self-efficacy. However, this methodology did not allow for a pre-post or a quasi-experimental design, and the effectiveness of our five-stage framework could not be confirmed quantitatively. In the future, conducting more experimental or design-based studies could further validate the effectiveness of our framework and broaden our findings. Furthermore, future studies should incorporate triangulation methods and utilize multiple data sources to enhance the reliability and validity of the findings. Meanwhile, owing to space limitations, we could only report the changes in Shuitao’s SE-IHT-IP and SE-IHT-SO here, and we could not describe the teacher self-efficacy of other participants regarding IAT. While nearly all of the pre-service teachers experienced an improvement in their teacher self-efficacy concerning IAT upon participating in our teacher professional development programs, the processes of their change were not entirely uniform. We will need to report the specific findings of these variations in the future. Further studies are also needed to explore the factors contributing to these variations. Moreover, following this study, we are implementing more teacher professional development programs of IAT. Future studies can explore the impact of this framework on additional aspects of pre-service STEAM teachers’ professional development. This will help gain a more comprehensive understanding of its effectiveness and potential areas for further improvement. Additionally, our five-stage framework was initially developed and implemented within the Chinese teacher education system. Future research should investigate how this framework can be adapted in other educational systems and cultural contexts.

The impetus behind this study stems from the burgeoning discourse advocating for the integration of A&H disciplines into STEM education on a global scale (e.g., Land 2020 ; Park and Cho 2022 ; Uştu et al. 2021 ; Vaziri and Bradburn 2021 ). Concurrently, there exists a pervasive concern regarding the challenges teachers face in implementing STEAM approaches, particularly in the context of IAT practices (e.g., Boice et al. 2021 ; Herro et al. 2019 ; Jacques et al. 2020 ; Park and Cho 2022 ; Perignat and Katz-Buonincontro 2019 ). To tackle this challenge, we first proposed a five-stage framework designed for teacher professional development programs of IAT. Then, utilizing this innovative framework, we implemented a series of teacher professional development programs. Drawing from the recommendations of Bray-Clark and Bates ( 2003 ), Kelley et al. ( 2020 ) and Zhou et al. ( 2023 ), we have selected teacher self-efficacy as a key metric to examine the effectiveness of the five-stage framework. Through a qualitative longitudinal case study, we scrutinized the influence of a specific teacher professional development program on the self-efficacy of a single pre-service teacher over an 18-month period. Our findings revealed a notable enhancement in teacher self-efficacy across both individual performance and student outcomes. The observed enhancement of the participant’s teacher self-efficacy did not occur swiftly but unfolded gradually through a prolonged, incremental process. Building on our findings, an updated model of teacher learning has been proposed. The updated model illustrates that teacher learning should be viewed as a continuous and sustainable process, wherein teaching performance, teacher beliefs, and teacher knowledge dynamically interact with one another. The updated model also confirms that teacher learning is inherently intertwined with student learning in STEAM education. Furthermore, this study also summarizes effective design features of STEAM teacher professional development programs.

Data availability

The datasets generated and/or analyzed during this study are not publicly available due to general data protection regulations, but are available from the corresponding author on reasonable request.

In their review article, Morris et al. ( 2017 ) equated “teaching self-efficacy” and “teacher self-efficacy” as synonymous concepts. This perspective is also adopted in this study.

An effective teacher professional development program should have specific, focused, and clear content instead of broad and scattered ones. Therefore, each pre-service teacher can only choose to integrate one subcomponent of A&H into their teaching in one teacher professional development program. For instance, Shuitao, a mathematics pre-service teacher, participated in one teacher professional development program focused on integrating history into mathematics teaching. However, she did not explore the integration of other subcomponents of A&H into her teaching during her graduate studies.

In the micro-classrooms, multi-angle, and multi-point high-definition video recorders are set up to record the teaching process.

In micro-teaching, mentors, in-service teachers, and other fellow pre-service teachers take on the roles of students.

In China, teachers can video record one section of a lesson and play them in formal classes. This is a practice known as a micro-course. For instance, in one teacher professional development program of integrating history into mathematics teaching, micro-courses encompass various mathematics concepts, methods, ideas, history-related material and related topics. Typically, teachers use these micro-courses to broaden students’ views, foster inquiry-based learning, and cultivate critical thinking skills. Such initiatives play an important role in improving teaching quality.

Many university-school collaboration initiatives in China focus on pre-service teachers’ practicum experiences (Lu et al. 2019 ). Our teacher professional development program is also supported by many K-12 schools in Shanghai. Personal information in videos is strictly protected.

In China, students are not required to pursue a graduate major that matches their undergraduate major. Most participants in our teacher professional development programs did not pursue undergraduate degrees in education-related fields.

Shuitao’s university reserves Wednesday afternoons for students to engage in various programs or clubs, as classes are not scheduled during this time. Similarly, our teacher professional development program activities are planned for Wednesday afternoons to avoid overlapping with participants’ other coursework commitments.

History is one of the most important components of A&H (Park and Cho 2022 ).

To learn more about genetic approach (i.e., genetic principle), see Jankvist ( 2009 ).

For the assessment process, see Fig. 2 .

Shuitao’s cooperating in-service teacher taught the concept of logarithms in Stage 2. In Stage 4, the teaching objective of her cooperating in-service teacher’s review lesson was to help students review the concept of logarithms to prepare students for the final exam.

Akiba M, Murata A, Howard C, Wilkinson B, Fabrega J (2019) Race to the top and lesson study implementation in Florida: District policy and leadership for teacher professional development. In: Huang R, Takahashi A, daPonte JP (eds) Theory and practice of lesson study in mathematics, pp. 731–754. Springer, Cham. https://doi.org/10.1007/978-3-030-04031-4_35

Alkhabra YA, Ibrahem UM, Alkhabra SA (2023) Augmented reality technology in enhancing learning retention and critical thinking according to STEAM program. Humanit Soc Sci Commun 10:174. https://doi.org/10.1057/s41599-023-01650-w

Article   Google Scholar  

Alwafi EM, Downey C, Kinchin G (2020) Promoting pre-service teachers’ engagement in an online professional learning community: Support from practitioners. J Professional Cap Community 5(2):129–146. https://doi.org/10.1108/JPCC-10-2019-0027

Archibald S, Coggshall JG, Croft A, Goe L (2011) High-quality professional development for all teachers: effectively allocating resources. National Comprehensive Center for Teacher Quality, Washington, DC

Google Scholar  

Bandura A (1977) Self-efficacy: Toward a unifying theory of behavioral change. Psychological Rev 84:191–215. https://doi.org/10.1016/0146-6402(78)90002-4

Article   CAS   Google Scholar  

Bandura A (1986) The explanatory and predictive scope of self-efficacy theory. J Soc Clin Psychol 4:359–373. https://doi.org/10.1521/jscp.1986.4.3.359

Bandura A (1997) Self-efficacy: The exercise of control. Freeman, New York

Basckin C, Strnadová I, Cumming TM (2021) Teacher beliefs about evidence-based practice: A systematic review. Int J Educ Res 106:101727. https://doi.org/10.1016/j.ijer.2020.101727

Bray-Clark N, Bates R (2003) Self-efficacy beliefs and teacher effectiveness: Implications for professional development. Prof Educ 26(1):13–22

Blonder R, Benny N, Jones MG (2014) Teaching self-efficacy of science teachers. In: Evans R, Luft J, Czerniak C, Pea C (eds), The role of science teachers’ beliefs in international classrooms: From teacher actions to student learning, Sense Publishers, Rotterdam, Zuid-Holland, pp. 3–16

Boice KL, Jackson JR, Alemdar M, Rao AE, Grossman S, Usselman M (2021) Supporting teachers on their STEAM journey: A collaborative STEAM teacher training program. Educ Sci 11(3):105. https://doi.org/10.3390/educsci11030105

Cai J (2017) Longitudinally investigating the effect of teacher professional development on instructional practice and student learning: A focus on mathematical problem posing. The University of Delaware, Newark, DE

Cai J, Chen T, Li X, Xu R, Zhang S, Hu Y, Zhang L, Song N (2020) Exploring the impact of a problem-posing workshop on elementary school mathematics teachers’ conceptions on problem posing and lesson design. Int J Educ Res 102:101404. https://doi.org/10.1016/j.ijer.2019.02.004

Calabrese JE, Capraro MM, Viruru R (2024) Semantic structure and problem posing: Preservice teachers’ experience. School Sci Math. https://doi.org/10.1111/ssm.12648

Clarke D, Hollingsworth H (2002) Elaborating a model of teacher professional growth. Teach Teach Educ 18(8):947–967. https://doi.org/10.1016/S0742-051X(02)00053-7

Corden A, Millar J (2007) Time and change: A review of the qualitative longitudinal research literature for social policy. Soc Policy Soc 6(4):583–592. https://doi.org/10.1017/S1474746407003910

Darling-Hammond L, Hyler ME, Gardner M (2017) Effective teacher professional development. Learning Policy Institute, Palo Alto, CA

Book   Google Scholar  

de la Garza A (2021) Internationalizing the curriculum for STEAM (STEM+ Arts and Humanities): From intercultural competence to cultural humility. J Stud Int Educ 25(2):123–135. https://doi.org/10.1177/1028315319888468

Article   MathSciNet   Google Scholar  

Desimone LM, Garet MS (2015) Best practices in teachers’ professional development in the United States. Psychol, Soc, Educ 7(3):252–263

Desimone LM, Porter AC, Garet MS, Yoon KS, Birman BF (2002) Effects of professional development on teachers’ instruction: Results from a three-year longitudinal study. Educ Evaluation Policy Anal 24(2):81–112. https://doi.org/10.3102/01623737024002081

Dicks SG, Northam HL, van Haren FM, Boer DP (2023) The bereavement experiences of families of potential organ donors: a qualitative longitudinal case study illuminating opportunities for family care. Int J Qualitative Stud Health Well-being 18(1):1–24. https://doi.org/10.1080/17482631.2022.2149100

Ding M, Huang R, Pressimone Beckowski C, Li X, Li Y (2024) A review of lesson study in mathematics education from 2015 to 2022: implementation and impact. ZDM Math Educ 56:87–99. https://doi.org/10.1007/s11858-023-01538-8

Duong NH, Nam NH, Trung TT (2024) Factors affecting the implementation of STEAM education among primary school teachers in various countries and Vietnamese educators: comparative analysis. Education 3–13. https://doi.org/10.1080/03004279.2024.2318239

English LD (2016) STEM education K-12: Perspectives on integration. Int J STEM Educ 3:3. https://doi.org/10.1186/s40594-016-0036-1

Garet MS, Porter AC, Desimone L, Birman BF, Yoon KS (2001) What makes professional development effective? Results from a national sample of teachers. Am Educ Res J 38(4):915–945. https://doi.org/10.3102/00028312038004915

Gates AE (2017) Benefits of a STEAM collaboration in Newark, New Jersey: Volcano simulation through a glass-making experience. J Geosci Educ 65(1):4–11. https://doi.org/10.5408/16-188.1

Geng J, Jong MSY, Chai CS (2019) Hong Kong teachers’ self-efficacy and concerns about STEM education. Asia-Pac Educ Researcher 28:35–45. https://doi.org/10.1007/s40299-018-0414-1

Han J, Kelley T, Knowles JG (2022) Building a sustainable model of integrated stem education: investigating secondary school STEM classes after an integrated STEM project. Int J Technol Design Educ. https://doi.org/10.1007/s10798-022-09777-8

Henderson S, Holland J, McGrellis S, Sharpe S, Thomson R (2012) Storying qualitative longitudinal research: Sequence, voice and motif. Qualitative Res 12(1):16–34. https://doi.org/10.1177/1468794111426232

Hennissen P, Beckers H, Moerkerke G (2017) Linking practice to theory in teacher education: A growth in cognitive structures. Teach Teach Educ 63:314–325. https://doi.org/10.1016/j.tate.2017.01.008

Henson RK (2002) From adolescent angst to adulthood: Substantive implications and measurement dilemmas in the development of teacher efficacy research. Educ Psychologist 37:137–150. https://doi.org/10.1207/S15326985EP3703_1

Herro D, Quigley C (2016) Innovating with STEAM in middle school classrooms: remixing education. Horizon 24(3):190–204. https://doi.org/10.1108/OTH-03-2016-0008

Herro D, Quigley C, Cian H (2019) The challenges of STEAM instruction: Lessons from the field. Action Teach Educ 41(2):172–190. https://doi.org/10.1080/01626620.2018.1551159

Huang B, Jong MSY, Tu YF, Hwang GJ, Chai CS, Jiang MYC (2022) Trends and exemplary practices of STEM teacher professional development programs in K-12 contexts: A systematic review of empirical studies. Comput Educ 189:104577. https://doi.org/10.1016/j.compedu.2022.104577

Hunter-Doniger T, Sydow L (2016) A journey from STEM to STEAM: A middle school case study. Clearing House 89(4-5):159–166. https://doi.org/10.1080/00098655.2016.1170461

Hwang S, Xu R, Yao Y, Cai J (2024) Learning to teach through problem posing: A teacher’s journey in a networked teacher−researcher partnership. J Math Behav 73:101120. https://doi.org/10.1016/j.jmathb.2023.101120

Jacques LA, Cian H, Herro DC, Quigley C (2020) The impact of questioning techniques on STEAM instruction. Action Teach Educ 42(3):290–308. https://doi.org/10.1080/01626620.2019.1638848

Jankvist UT (2009) A categorization of the “whys” and “hows” of using history in mathematics education. Educ Stud Math 71:235–261. https://doi.org/10.1007/s10649-008-9174-9

Jiang H, Chugh R, Turnbull D, Wang X, Chen S (2023) Modeling the impact of intrinsic coding interest on STEM career interest: evidence from senior high school students in two large Chinese cities. Educ Inf Technol 28:2639–2659. https://doi.org/10.1007/s10639-022-11277-0

Jiang H, Chugh R, Turnbull D, Wang X, Chen S (2024a) Exploring the effects of technology-related informal mathematics learning activities: A structural equation modeling analysis. Int J Sci Math Educ . Advance online publication. https://doi.org/10.1007/s10763-024-10456-4

Jiang H, Islam AYMA, Gu X, Guan J (2024b) How do thinking styles and STEM attitudes have effects on computational thinking? A structural equation modeling analysis. J Res Sci Teach 61:645–673. https://doi.org/10.1002/tea.21899

Jiang H, Turnbull D, Wang X, Chugh R, Dou Y, Chen S (2022) How do mathematics interest and self-efficacy influence coding interest and self-efficacy? A structural equation modeling analysis. Int J Educ Res 115:102058. https://doi.org/10.1016/j.ijer.2022.102058

Jiang H, Wang K, Wang X, Lei X, Huang Z (2021) Understanding a STEM teacher’s emotions and professional identities: A three-year longitudinal case study. Int J STEM Educ 8:51. https://doi.org/10.1186/s40594-021-00309-9

Kelley TR, Knowles JG, Holland JD, Han J (2020) Increasing high school teachers self-efficacy for integrated STEM instruction through a collaborative community of practice. Int J STEM Educ 7:14. https://doi.org/10.1186/s40594-020-00211-w

Klassen RM, Tze VMC, Betts SM, Gordon KA (2011) Teacher efficacy research 1998–2009: Signs of progress or unfulfilled promise? Educ Psychol Rev 23(1):21–43. https://doi.org/10.1007/s10648-010-9141-8

Land M (2020) The importance of integrating the arts into STEM curriculum. In: Stewart AJ, Mueller MP, Tippins DJ (eds), Converting STEM into STEAM programs, pp. 11–19. Springer. https://doi.org/10.1007/978-3-030-25101-7_2

Lee C, Bobko P (1994) Self-efficacy beliefs: Comparison of five measures. J Appl Psychol 79(3):364–369. https://doi.org/10.1037/0021-9010.79.3.364

Li W, Chiang FK (2019) Preservice teachers’ perceptions of STEAM education and attitudes toward STEAM disciplines and careers in China. In: Sengupta P, Shanahan MC, Kim B, (eds), Critical, transdisciplinary and embodied approaches in STEM education. Springer. https://doi.org/10.1007/978-3-030-29489-2_5

Liu M, Zwart R, Bronkhorst L, Wubbels T (2022) Chinese student teachers’ beliefs and the role of teaching experiences in the development of their beliefs. Teach Teach Educ 109:103525. https://doi.org/10.1016/j.tate.2021.103525

Liu S, Xu X, Stronge J (2018) The influences of teachers’ perceptions of using student achievement data in evaluation and their self-efficacy on job satisfaction: evidence from China. Asia Pac Educ Rev 19:493–509. https://doi.org/10.1007/s12564-018-9552-7

Long T, Zhao G, Yang X, Zhao R, Chen Q (2021) Bridging the belief-action gap in a teachers’ professional learning community on teaching of thinking. Professional Dev Educ 47(5):729–744. https://doi.org/10.1080/19415257.2019.1647872

Lu L, Wang F, Ma Y, Clarke A, Collins J (2019) Exploring Chinese teachers’ commitment to being a cooperating teacher in a university-government-school initiative for rural practicum placements. In: Liu WC, Goh CCM (eds), Teachers’ perceptions, experience and learning, pp. 33–54. Routledge, London

Lyu S, Niu S, Yuan J, Zhan Z (2024) Developing professional capital through technology-enabled university-school-enterprise collaboration: an innovative model for C-STEAM preservice teacher education in the Greater Bay area. Asia Pacific J Innov Entrepreneurship. https://doi.org/10.1108/APJIE-01-2024-0014

Marco N, Palatnik A (2024) From “learning to variate” to “variate for learning”: Teachers learning through collaborative, iterative context-based mathematical problem posing. J Math Behav 73:101119. https://doi.org/10.1016/j.jmathb.2023.101119

Merriam SB (1998) Qualitative research and case study applications in education. Jossey-Bass Publishers, Hoboken, New Jersey

Morris DB, Usher EL (2011) Developing teaching self-efficacy in research institutions: A study of award-winning professors. Contemp Educ Psychol 36(3):232–245. https://doi.org/10.1016/j.cedpsych.2010.10.005

Morris DB, Usher EL, Chen JA (2017) Reconceptualizing the sources of teaching self-efficacy: A critical review of emerging literature. Educ Psychol Rev 29(4):795–833. https://doi.org/10.1007/s10648-016-9378-y

Matusovich HM, Streveler RA, Miller RL (2010) Why do students choose engineering? A qualitative, longitudinal investigation of students’ motivational values. J Eng Educ 99(4):289–303. https://doi.org/10.1002/j.2168-9830.2010.tb01064.x

Näykki P, Kontturi H, Seppänen V, Impiö N, Järvelä S (2021) Teachers as learners–a qualitative exploration of pre-service and in-service teachers’ continuous learning community OpenDigi. J Educ Teach 47(4):495–512. https://doi.org/10.1080/02607476.2021.1904777

OECD (2018) Teaching and learning international survey (TALIS) 2018 conceptual framework. OECD, Paris

Örnek T, Soylu Y (2021) A model design to be used in teaching problem posing to develop problem-posing skills. Think Skills Creativity 41:100905. https://doi.org/10.1016/j.tsc.2021.100905

Ord K, Nuttall J (2016) Bodies of knowledge: The concept of embodiment as an alternative to theory/practice debates in the preparation of teachers. Teach Teach Educ 60:355–362. https://doi.org/10.1016/j.tate.2016.05.019

Ozkan G, Umdu Topsakal U (2021) Investigating the effectiveness of STEAM education on students’ conceptual understanding of force and energy topics. Res Sci Technol Educ 39(4):441–460. https://doi.org/10.1080/02635143.2020.1769586

Park W, Cho H (2022) The interaction of history and STEM learning goals in teacher-developed curriculum materials: opportunities and challenges for STEAM education. Asia Pacific Educ Rev. https://doi.org/10.1007/s12564-022-09741-0

Perignat E, Katz-Buonincontro J (2019) STEAM in practice and research: An integrative literature review. Think Skills Creativity 31:31–43. https://doi.org/10.1016/j.tsc.2018.10.002

Philipp RA (2007) Mathematics teachers’ beliefs and affect. In: Lester Jr FK, (ed), Second handbook of research on mathematics teaching and learning, pp. 257–315. Information Age, Charlotte, NC

Quigley CF, Herro D, Jamil FM (2017) Developing a conceptual model of STEAM teaching practices. Sch Sci Math 117(1–2):1–12. https://doi.org/10.1111/ssm.12201

Ro S, Xiao S, Zhou Z (2022) Starting up STEAM in China: A case study of technology entrepreneurship for STEAM education in China. In: Ray P, Shaw R (eds), Technology entrepreneurship and sustainable development, pp. 115–136. Springer. https://doi.org/10.1007/978-981-19-2053-0_6

Roth KJ, Bintz J, Wickler NI, Hvidsten C, Taylor J, Beardsley PM, Wilson CD (2017) Design principles for effective video-based professional development. Int J STEM Educ 4:31. https://doi.org/10.1186/s40594-017-0091-2

Article   PubMed   PubMed Central   Google Scholar  

Ryu M, Mentzer N, Knobloch N (2019) Preservice teachers’ experiences of STEM integration: Challenges and implications for integrated STEM teacher preparation. Int J Technol Des Educ, 29:493–512. https://doi.org/10.1007/s10798-018-9440-9

Sancar R, Atal D, Deryakulu D (2021) A new framework for teachers’ professional development. Teach Teach Educ 101:103305. https://doi.org/10.1016/j.tate.2021.103305

Sanz-Camarero R, Ortiz-Revilla J, Greca IM (2023) The place of the arts within integrated education. Arts Educ Policy Rev. https://doi.org/10.1080/10632913.2023.2260917

Shirani F, Henwood K (2011) Continuity and change in a qualitative longitudinal study of fatherhood: relevance without responsibility. Int J Soc Res Methodol 14(1):17–29. https://doi.org/10.1080/13645571003690876

Smith CE, Paré JN (2016) Exploring Klein bottles through pottery: A STEAM investigation. Math Teach 110(3):208–214. https://doi.org/10.5951/mathteacher.110.3.0208

Soprano K, Yang L (2013) Inquiring into my science teaching through action research: A case study on one pre-service teacher’s inquiry-based science teaching and self-efficacy. Int J Sci Math Educ 11(6):1351–1368. https://doi.org/10.1007/s10763-012-9380-x

Stake RE (1995) The art of case study research. Sage Publication, Thousand Oaks, California

Stohlmann M, Moore T, Roehrig G (2012) Considerations for teaching integrated STEM education. J Pre Coll Eng Educ Res 2(1):28–34. https://doi.org/10.5703/1288284314653

Strauss AL, Corbin JM (1990) Basics of qualitative research. Sage Publications, Thousand Oaks, California

Taimalu M, Luik P (2019) The impact of beliefs and knowledge on the integration of technology among teacher educators: A path analysis. Teach Teach Educ 79:101–110. https://doi.org/10.1016/j.tate.2018.12.012

Takahashi A, McDougal T (2016) Collaborative lesson research: Maximizing the impact of lesson study. ZDM Math Educ 48:513–526. https://doi.org/10.1007/s11858-015-0752-x

Thompson AG (1992) Teachers’ beliefs and conceptions: A synthesis of the research. In: Grouws DA (ed), Handbook of research on mathematics teaching and learning, pp. 127–146. Macmillan, New York

Tschannen-Moran M, Woolfolk Hoy A, Hoy WK (1998) Teacher efficacy: Its meaning and measure. Rev Educ Res 68:202–248. https://doi.org/10.3102/00346543068002202

Unfried A, Rachmatullah A, Alexander A, Wiebe E (2022) An alternative to STEBI-A: validation of the T-STEM science scale. Int J STEM Educ 9:24. https://doi.org/10.1186/s40594-022-00339-x

Uştu H, Saito T, Mentiş Taş A (2021) Integration of art into STEM education at primary schools: an action research study with primary school teachers. Syst Pract Action Res. https://doi.org/10.1007/s11213-021-09570-z

Vaziri H, Bradburn NM (2021) Flourishing effects of integrating the arts and humanities in STEM Education: A review of past studies and an agenda for future research. In: Tay L, Pawelski JO (eds), The Oxford handbook of the positive humanities, pp. 71–82. Oxford University Press, Oxford

Wang K, Wang X, Li Y, Rugh MS (2018) A framework for integrating the history of mathematics into teaching in Shanghai. Educ Stud Math 98:135–155. https://doi.org/10.1007/s10649-018-9811-x

Wang Z, Jiang H, Jin W, Jiang J, Liu J, Guan J, Liu Y, Bin E (2024) Examining the antecedents of novice stem teachers’ job satisfaction: The roles of personality traits, perceived social support, and work engagement. Behav Sci 14(3):214. https://doi.org/10.3390/bs14030214

Wenger E, McDermott R, Snyder WM (2002) Cultivating communities of practice. Harvard Business School Press, Boston, MA

Wong JT, Bui NN, Fields DT, Hughes BS (2022) A learning experience design approach to online professional development for teaching science through the arts: Evaluation of teacher content knowledge, self-efficacy and STEAM perceptions. J Sci Teacher Educ. https://doi.org/10.1080/1046560X.2022.2112552

Wyatt M (2015) Using qualitative research methods to assess the degree of fit between teachers’ reported self-efficacy beliefs and their practical knowledge during teacher education. Aust J Teach Educ 40(1):117–145

Xenofontos C, Andrews P (2020) The discursive construction of mathematics teacher self-efficacy. Educ Stud Math 105(2):261–283. https://doi.org/10.1007/s10649-020-09990-z

Yin R (2003) Case study research: Design and methods. Sage Publications, Thousand Oaks, California

Zakariya YF (2020) Effects of school climate and teacher self-efficacy on job satisfaction of mostly STEM teachers: a structural multigroup invariance approach. Int J STEM Educ 7:10. https://doi.org/10.1186/s40594-020-00209-4

Zee M, de Jong PF, Koomen HM (2017) From externalizing student behavior to student-specific teacher self-efficacy: The role of teacher-perceived conflict and closeness in the student–teacher relationship. Contemp Educ Psychol 51:37–50. https://doi.org/10.1016/j.cedpsych.2017.06.009

Zee M, Koomen HM (2016) Teacher self-efficacy and its effects on classroom processes, student academic adjustment, and teacher well-being: A synthesis of 40 years of research. Rev Educ Res 86(4):981–1015. https://doi.org/10.3102/0034654315626801

Zhan Z, Li Y, Mei H, Lyu S (2023) Key competencies acquired from STEM education: gender-differentiated parental expectations. Humanit Soc Sci Commun 10:464. https://doi.org/10.1057/s41599-023-01946-x

Zhan Z, Niu S (2023) Subject integration and theme evolution of STEM education in K-12 and higher education research. Humanit Soc Sci Commun 10:781. https://doi.org/10.1057/s41599-023-02303-8

Zhong B, Liu X, Li X (2024) Effects of reverse engineering pedagogy on students’ learning performance in STEM education: The bridge-design project as an example. Heliyon 10(2):e24278. https://doi.org/10.1016/j.heliyon.2024.e24278

Zhong B, Liu X, Zhan Z, Ke Q, Wang F (2022) What should a Chinese top-level design in STEM Education look like? Humanit Soc Sci Commun 9:261. https://doi.org/10.1057/s41599-022-01279-1

Zhou B (2021) Cultivation of Ed. M. to bring up “famous subject teachers”: practical exploration and policy suggestions. Teach Educ Res 33(5):19–26. https://doi.org/10.13445/j.cnki.t.e.r.2021.05.003

Zhou X, Shu L, Xu Z, Padrón Y (2023) The effect of professional development on in-service STEM teachers’ self-efficacy: a meta-analysis of experimental studies. Int J STEM Educ 10:37. https://doi.org/10.1186/s40594-023-00422-x

Download references

Acknowledgements

This research is funded by 2021 National Natural Science Foundation of China (Grant No.62177042), 2024 Zhejiang Provincial Natural Science Foundation of China (Grant No. Y24F020039), and 2024 Zhejiang Educational Science Planning Project (Grant No. 2024SCG247).

Author information

Xuesong Zhai

Present address: School of Education, City University of Macau, Macau, China

Authors and Affiliations

College of Education, Zhejiang University, Hangzhou, China

Haozhe Jiang & Xuesong Zhai

School of Engineering and Technology, CML‑NET & CREATE Research Centres, Central Queensland University, North Rockhampton, QLD, Australia

Ritesh Chugh

Hangzhou International Urbanology Research Center & Zhejiang Urban Governance Studies Center, Hangzhou, China

Department of Teacher Education, Nicholls State University, Thibodaux, LA, USA

School of Mathematical Sciences, East China Normal University, Shanghai, China

Xiaoqin Wang

College of Teacher Education, Faculty of Education, East China Normal University, Shanghai, China

You can also search for this author in PubMed   Google Scholar

Contributions

Conceptualization - Haozhe Jiang; methodology - Haozhe Jiang; software - Xuesong Zhai; formal analysis - Haozhe Jiang & Ke Wang; investigation - Haozhe Jiang; resources - Haozhe Jiang, Xuesong Zhai & Xiaoqin Wang; data curation - Haozhe Jiang & Ke Wang; writing—original draft preparation - Haozhe Jiang & Ritesh Chugh; writing—review and editing - Ritesh Chugh & Ke Wang; visualization - Haozhe Jiang, Ke Wang & Xiaoqin Wang; supervision - Xuesong Zhai & Xiaoqin Wang; project administration - Xuesong Zhai & Xiaoqin Wang; and funding acquisition - Xuesong Zhai & Xiaoqin Wang. All authors have read and agreed to the published version of the manuscript.

Corresponding author

Correspondence to Xuesong Zhai .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Ethical approval

This research was approved by the Committee for Human Research of East China Normal University (Number: HR 347-2022). The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

Informed consent

Written informed consent was obtained from all participants in this study before data collection.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Jiang, H., Chugh, R., Zhai, X. et al. Longitudinal analysis of teacher self-efficacy evolution during a STEAM professional development program: a qualitative case study. Humanit Soc Sci Commun 11 , 1162 (2024). https://doi.org/10.1057/s41599-024-03655-5

Download citation

Received : 27 April 2024

Accepted : 12 August 2024

Published : 08 September 2024

DOI : https://doi.org/10.1057/s41599-024-03655-5

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

How pre-service teacher self-efficacy changes during the professional experience placement: a growth mixture model

• Published: 06 September 2024

Cite this article

• Kang Ma   ORCID: orcid.org/0000-0002-2600-7150 1 ,
• Jingjing Dong 1 ,
• Anne McMaugh   ORCID: orcid.org/0000-0003-2988-0366 2 &
• Youxing Cui 3  

18 Accesses

Explore all metrics

Longitudinal studies are essential to examine teacher self-efficacy (TSE) changes during the formative stages of pre-service teacher development. However, limited longitudinal studies have been conducted on this population. Among the existing studies, no study has been found to have applied an individual-centred modelling approach to examine TSE changes; similarly, few studies have examined longitudinal change in domains other than classroom teaching, nor in the Chinese context. The present study surveyed 205 pre-service teachers at three time points, namely at the beginning, middle and end of the professional experience placement. This study used a two-domain TSE scale assessing classroom teaching and school-level activities. The data were analysed using growth mixture modelling in Mplus. Three types of trajectories were fit for TSE for classroom teaching and five types were fit for school-level activities. For classroom teaching, trajectories reflect a high-TSE increasing class, low-TSE decreasing class, and medium-TSE non-increasing class. For school-level activities, trajectories reflect high-TSE increasing, medium-TSE non-increasing class, low-TSE greater increasing class, high-TSE decreasing class, and low-TSE non-increasing class. The classification of TSE trajectories differed according to gender and degree program. Implications, limitations, and suggestions for future research are discussed.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Similar content being viewed by others

Transitioning into the first year of teaching: changes and sources of teacher self-efficacy

Discovering the impact of reading coursework and discipline-specific mentorship on first-year teachers’ self-efficacy: a latent class analysis

Identifying teachers’ motivational profiles and their changes from teacher education into practice: A longitudinal study

Avanzi, L., Miglioretti, M., Velasco, V., Dalducci, C., Vecchio, L., Fraccaroli, F., & Skaalvik, E. M. (2013). Cross-validation of the Norwegian Teacher’s Self-Efficacy Scale (NTSES). Teaching and Teacher Education, 31 (1), 69–78. https://doi.org/10.1016/j.tate.2013.01.002

Article   Google Scholar  

Baba, K., & Nitta, R. (2021). Emergence of multiple groups of learners with different writing-development trajectories in classroom: Growth mixture modeling. Journal of Second Language Writing, 54 (December), 100856. https://doi.org/10.1016/j.jslw.2021.100856

Bandura, A. (1997). Self-efficacy: The exercise of control. Freeman .

Bandura, A. (2006). Guide for creating self-efficacy scales. In F. Pajares & T. Urdan (Eds.), Self-efficacy beliefs of adolescents (Vol. 5, pp. 307–337). Information Age.

Google Scholar  

Bandura, A. (2012). On the functional properties of perceived self-efficacy revisited. Journal of Management, 38 (1), 9–44. https://doi.org/10.1177/0149206311410606

bin Khairani, A. Z., & bin Abd Razak, N. (2012). An analysis of the Teachers’ Sense of Efficacy Scale within the Malaysian context using the Rasch measurement mode. Procedia: Social and Behavioral Sciences, 69 , 2137–2142. https://doi.org/10.1016/j.sbspro.2012.12.178

Boeve-de Pauw, J., Olsson, D., Berglund, T., & Gericke, N. (2022). Teachers’ ESD self-efficacy and practices: A longitudinal study on the impact of teacher professional development. Environmental Education Research, 28 (6), 867–885. https://doi.org/10.1080/13504622.2022.2042206

Burgueño, R., Sicilia, A., Medina-Casaubón, J., Alcaraz-Ibañez, M., & Lirola, M. J. (2018). Psychometry of the Teacher’s Sense of Efficacy Scale in Spanish teachers’ education. The Journal of Experimental Education, 87 (1), 89–100. https://doi.org/10.1080/00220973.2018.1423542

Chang, Y. L. (2010). A case study of elementary beginning mathematics teachers’ efficacy development. International Journal of Science and Mathematics Education, 8 (2), 271–297. https://doi.org/10.1007/s10763-009-9173-z

Charalambous, C. Y., Philippou, G. N., & Kyriakides, L. (2008). Tracing the development of pre-service teachers’ efficacy beliefs in teaching mathematics during fieldwork. Educational Studies in Mathematics, 67 (2), 125–142. https://doi.org/10.1007/s10649-007-9084-2

Cheong, D. (2010). The effects of practice teaching sessions in second life on the change in pre-service teachers’ teaching efficacy. Computers and Education, 55 (2), 868–880. https://doi.org/10.1016/j.compedu.2010.03.018

Cherniss, C. (1993). Role of professional self-efficacy in the etiology and amelioration of burnout. In Professional burnout: Recent developments in theory and research. (pp. 135–149). Taylor & Francis.

Chesnut, S. R., & Burley, H. (2015). Self-efficacy as a predictor of commitment to the teaching profession: A meta-analysis. Educational Research Review, 15 , 1–16. https://doi.org/10.1016/j.edurev.2015.02.001

Cocca, M., & Cocca, A. (2021). Testing a four-factor model for the Teachers’ Sense of Efficacy Scale: An updated perspective on teachers’ perceived classroom efficacy. Psicología Educativa, 28 (1), 39–46. https://doi.org/10.5093/psed2021a3

De Neve, D., & Devos, G. (2017). Psychological states and working conditions buffer beginning teachers’ intention to leave the job. European Journal of Teacher Education, 40 (1), 6–27. https://doi.org/10.1080/02619768.2016.1246530

Enders, C. K. (2013). Dealing with missing data in developmental research. Child Development Perspectives, 7 (1), 27–31. https://doi.org/10.1111/cdep.12008

Fackler, S., & Malmberg, L. E. (2016). Teachers’ self-efficacy in 14 OECD countries: Teacher, student group, school and leadership effects. Teaching and Teacher Education, 56 , 185–195. https://doi.org/10.1016/j.tate.2016.03.002

Feng, L., Hodges, T. S., Waxman, H. C., & Joshi, R. M. (2019). Discovering the impact of reading coursework and discipline-specific mentorship on first-year teachers’ self-efficacy: A latent class analysis. Annals of Dyslexia, 69 (1), 80–98. https://doi.org/10.1007/s11881-018-00167-1

Fitchett, P. G., McCarthy, C. J., Lambert, R. G., & Boyle, L. (2018). An examination of US first-year teachers’ risk for occupational stress: Associations with professional preparation and occupational health. Teachers and Teaching: Theory and Practice, 24 (2), 99–118. https://doi.org/10.1080/13540602.2017.1386648

Fives, H., & Buehl, M. M. (2009). Examining the factor structure of the teachers’ sense of efficacy scale. Journal of Experimental Education, 78 (1), 118–134. https://doi.org/10.1080/00220970903224461

Friedman, I. A., & Kass, E. (2002). Teacher self-efficacy: A classroom-organization conceptualization. Teaching & Teacher Education, 18 (6), 675.

Ginns, I. S., Tulip, D. F., Watters, J. J., & Lucas, K. B. (1995). Changes in preservice elementary teachers’ sense of efficacy in teaching science. School Science and Mathematics, 95 (8), 394–400. https://doi.org/10.1111/j.1949-8594.1995.tb10191.x

Gurvitch, R., & Metzler, M. W. (2009). The effects of laboratory-based and field-based practicum experience on pre-service teachers’ self-efficacy. Teaching and Teacher Education, 25 (3), 437–443. https://doi.org/10.1016/j.tate.2008.08.006

Harlin, J. F., Roberts, T., Mowen, G., & Edgar, D. W. (2007). A longitudinal examination of teaching efficacy of agricultural science student teachers at four different institutions. Journal of Agricultural Education, 48 (3), 78–90.

Harrison, X. A., Donaldson, L., Correa-Cano, M. E., Evans, J., Fisher, D. N., Goodwin, C. D., Robinson, B. S., Hodgson, D. D., & Inger, R. (2018). A brief introduction to mixed effects modelling and multi-model inference in ecology. PeerJ, 6 , 1–32. https://doi.org/10.7717/peerj.4794

Henson, R. K. (2001). Teacher self-efficacy: Substantive implications and measurement dilemmas. Annual Meeting of the Educational Research Exchange , March , 24. https://files.eric.ed.gov/fulltext/ED452208.pdf

Hettinger, K., Lazarides, R., Rubach, C., & Schifele, U. (2021). Teacher classroom management self-efficacy: Longitudinal relations to perceived teaching behaviors and student enjoyment. Teaching and Teacher Education, 103 , 103349. https://doi.org/10.1016/j.tate.2021.103349

Hoang, T., & Wyatt, M. (2021). Exploring the self-efficacy beliefs of Vietnamese pre-service teachers of English as a foreign language. System, 96 , 102422. https://doi.org/10.1016/j.system.2020.102422

Huang, L., Zhang, T., & Huang, Y. (2020). Effects of school organizational conditions on teacher professional learning in China: The mediating role of teacher self-efficacy. Studies in Educational Evaluation, 66 , 100893. https://doi.org/10.1016/j.stueduc.2020.100893

Jackson, N., & Miller, R. (2020). Teacher candidates’ sense of self-efficacy toward classroom management. Journal of Education, 200 (3), 153–163. https://doi.org/10.1177/0022057419881169

Johnson, S. K. (2021). Latent profile transition analyses and growth mixture models: A very non-technical guide for researchers in child and adolescent development. New Directions for Child and Adolescent Development, 175 , 111–139. https://doi.org/10.1002/cad.20398

Kim, L. E., Jörg, V., & Klassen, R. M. (2019). A meta-analysis of the effects of teacher personality on teacher effectiveness and burnout. Educational Psychology Review, 31 (1), 163–195. https://doi.org/10.1007/s10648-018-9458-2

Klassen, R. M., & Durksen, T. L. (2014). Weekly self-efficacy and work stress during the teaching practicum: A mixed methods study. Learning & Instruction, 33 , 158–169.

Klassen, R. M., & Tze, V. M. C. (2014). Teachers’ self-efficacy, personality, and teaching effectiveness: A meta-analysis. Educational Research Review, 12 , 59–76. https://doi.org/10.1016/j.edurev.2014.06.001

Koniewski, M. (2019). The teacher self-efficacy scale (TSES) factorial structure evidence review and new evidence from polish-speaking samples. European Journal of Psychological Assessment, 35 (6), 900–912. https://doi.org/10.1027/1015-5759/a000475

Ma, K., Trevethan, R., & Lu, S. (2019). Measuring teacher sense of efficacy: Insights and recommendations concerning scale design and data analysis from research with preservice and inservice teachers in china. Frontiers of Education in China, 14 , 612–686. https://doi.org/10.1007/s11516-019-0029-1

Ma, K., Cavanagh, M. S., & McMaugh, A. (2021). Preservice teachers’ reflections on their teaching self-efficacy changes for the first professional experience placement. Australian Journal of Teacher Education, 46 (10), 62. https://doi.org/10.14221/ajte.2021v46n10.4

Ma, K., McMaugh, A., & Cavanagh, M. (2021). The development of teacher self-efficacy from preservice to early career teacher: A systematic review of development and methodological quality in longitudinal research. International Journal of Research & Method in Education, 45 (5), 450–465. https://doi.org/10.1080/1743727X.2021.1990879

Ma, K., Luo, J., Cavanagh, M., Dong, J., & Sun, M. (2023). Measuring teacher self-efficacy: Validating a new comprehensive scale among Chinese pre-service teachers. Frontiers in Psychology., 13 , 1063830. https://doi.org/10.3389/fpsyg.2022.1063830

Marschall, G. (2021). The role of teacher identity in teacher self-efficacy development: The case of Katie. Journal of Mathematics Teacher Education, 25 , 725–747. https://doi.org/10.1007/s10857-021-09515-2

Mok, M. M. C., & Moore, P. J. (2019). Teachers & self-efficacy. Educational Psychology, 39 (1), 1–3. https://doi.org/10.1080/01443410.2019.1567070

O’Neill, S. C. (2016). Preparing preservice teachers for inclusive classrooms: Does completing coursework on managing challenging behaviours Increase their classroom management sense of efficacy? Australasian Journal of Special Education, 40 (2), 117–140. https://doi.org/10.1017/jse.2015.10

Pajares, F., & Johnson, M. J. (1994). Confidence and competence in Writing: The role of self-efficacy, and apprehension. Research in the Teaching of English, 28 (3), 313–331.

Perera, H. N., Calkins, C., & Part, R. (2019). Teacher self-efficacy profiles: Determinants, outcomes, and generalizability across teaching level. Contemporary Educational Psychology, 58 , 186–203. https://doi.org/10.1016/j.cedpsych.2019.02.006

Petersen, K. J., Qualter, P., & Humphrey, N. (2019). The application of latent class analysis for investigating population child mental health: A systematic review. Frontiers in Psychology, 10 , 01214. https://doi.org/10.3389/fpsyg.2019.01214

Pfitzner-Eden, F. (2016). I feel less confident so I quit? Do true changes in teacher self-efficacy predict changes in pre-service teachers’ intention to quit their teaching degree? Teaching and Teacher Education, 55 , 240–254. https://doi.org/10.1016/j.tate.2016.01.018

Preacher, K. J., & MacCallum, R. C. (2003). Repairing Tom Swift’s electric factor analysis machine. Understanding Statistics, 2 (1), 13–43. https://doi.org/10.1207/s15328031us0201_02

Ram, N., & Grimm, K. J. (2009). Methods and measures: Growth mixture modeling: A method for identifying differences in longitudinal change among unobserved groups. International Journal of Behavioral Development, 33 (6), 565–576. https://doi.org/10.1177/0165025409343765

Raymond, S., & Gabriel, F. (2023). An ecological framework for early years teacher self-efficacy development. Teaching and Teacher Education, 132 , 104252. https://doi.org/10.1016/j.tate.2023.104252

Şahin, F. E., & Atay, D. (2010). Sense of efficacy from student teaching to the induction year. Procedia: Social and Behavioral Sciences, 2 (2), 337–341. https://doi.org/10.1016/j.sbspro.2010.03.021

Salas-Rodríguez, F., Lara, S., & Martínez, M. (2021). Spanish version of the Teachers’ Sense of Efficacy Scale: An adaptation and validation study. Frontiers in Psychology, 12 , 1–12. https://doi.org/10.3389/fpsyg.2021.714145

Sciuchetti, M. B., & Yssel, N. (2019). The development of pre-service teachers’ self- efficacy for classroom and behavior management across multiple field experiences. Australian Journal of Teacher Education, 44 (6), 18–34. https://doi.org/10.14221/ajte.2018v44n6.2

Siwatu, K. O. (2011). Pre-service teachers’ sense of preparedness and self-efficacy to teach in America’s urban and suburban schools: Does context matter? Teaching and Teacher Education, 27 (2), 357–365. https://doi.org/10.1016/j.tate.2010.09.004

Skaalvik, E. M., & Skaalvik, S. (2007). Dimensions of teacher self-efficacy and relations with strain factors, perceived collective teacher efficacy, and teacher burnout. Journal of Educational Psychology, 99 (3), 611–625.

Symes, W., Lazarides, R., & Hußner, I. (2023). The development of student teachers’ teacher self-efficacy before and during the COVID-19 pandemic. Teaching and Teacher Education, 122 , 103941. https://doi.org/10.1016/j.tate.2022.103941

Thomson, M. M., Huggins, E., & Williams, W. (2019). Developmental science efficacy trajectories of novice teachers from a STEM-Focused program: A longitudinal mixed-methods investigation. Teaching and Teacher Education, 77 , 253–265. https://doi.org/10.1016/j.tate.2018.10.010

Tschannen-Moran, M., & Hoy, A. W. (2001). Teacher efficacy: Capturing an elusive construct. Teaching and Teacher Education, 17 (7), 783–805. https://doi.org/10.1016/S0742-051X(01)00036-1

Valls, M., Bonvin, P., & Benoit, V. (2020). Psychometric properties of the French version of the Teachers’ Sense of Efficacy Scale (TSES-12f). Revue Europeenne De Psychologie Appliquee, 70 (3), 100551. https://doi.org/10.1016/j.erap.2020.100551

Wang, Z., Oh, W., Malanchini, M., & Borriello, G. (2020). The developmental trajectories of mathematics anxiety: Cognitive, personality, and environmental correlates. Contemporary Educational Psychology, 61 , 101876. https://doi.org/10.1016/j.cedpsych.2020.101876

Weller, B. E., Bowen, N. K., & Faubert, S. J. (2020). Latent class analysis: A guide to best practice. Journal of Black Psychology, 46 (4), 287–311. https://doi.org/10.1177/0095798420930932

Wickrama, K., Lee, T. K., O’Neal, C. W., & Lorenz, F. (2021). Higher-order growth curves and mixture modeling with Mplus: A practical guide (2nd ed.). Routledge.

Book   Google Scholar  

Woolfolk Hoy, A., & Spero, R. B. (2005). Changes in teacher efficacy during the early years of teaching: A comparison of four measures. Teaching and Teacher Education, 21 , 343–356. https://doi.org/10.1016/j.tate.2005.06.010

Yada, A., Leskinen, M., Savolainen, H., & Schwab, S. (2022). Meta-analysis of the relationship between teachers’ self-efficacy and attitudes toward inclusive education. Teaching and Teacher Education, 109 , 1–15. https://doi.org/10.1016/j.tate.2021.103521

Zee, M., & Koomen, H. M. (2016). Teacher self-efficacy and its effects on classroom processes, student academic adjustment, and teacher well-being. Review of Educational Research, 86 (4), 981–1015. https://doi.org/10.3102/0034654315626801

Zhu, G., Mena, J., & Johnson, C. (2020). Chinese student teachers’ teaching practicum experiences: Insights from transformative learning, third space, and dialogical-self theory. International Journal of Educational Research, 103 , 101638. https://doi.org/10.1016/j.ijer.2020.101638

Download references

This work was supported by Jiangsu Provincial Funding for Social Science [23JYD014].

Author information

Authors and affiliations.

School of Education, Yancheng Teachers University, Yancheng, China

Kang Ma & Jingjing Dong

School of Education, Macquarie University, Sydney, Australia

Anne McMaugh

School of Education, Hainan Normal University, Haikou, China

Youxing Cui

You can also search for this author in PubMed   Google Scholar

Corresponding authors

Correspondence to Kang Ma or Youxing Cui .

Ethics declarations

Conflict of interest.

The authors declare that there is no conflict of interests

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Ma, K., Dong, J., McMaugh, A. et al. How pre-service teacher self-efficacy changes during the professional experience placement: a growth mixture model. Aust. Educ. Res. (2024). https://doi.org/10.1007/s13384-024-00766-5

Download citation

Received : 11 January 2024

Accepted : 14 August 2024

Published : 06 September 2024

DOI : https://doi.org/10.1007/s13384-024-00766-5

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Pre-service teacher education
• Teacher self-efficacy
• Professional experience placement
• Longitudinal studies
• Growth mixture modelling
• Find a journal
• Publish with us
• Track your research

• Our Mission

Teacher Development Research Review: Keys to Educator Success

How can you get the best out of your teachers and improve student learning? Edutopia’s research analyst explains some of the best practices found by researchers to help ensure educator growth and success

Teaching quality has been defined as "instruction that enables a wide range of students to learn" ( Darling-Hammond, 2012 ), and it is the strongest school-related factor that can improve student learning and achievement ( Hanushek, 2011 ; Nye, Konstantopoulos, and Hedges, 2004 ; Rivkin, Hanushek, and Kain, 2005 ). Knowing this, what is the best way to foster and provide ongoing support for good teaching practices? While every school is unique, research has identified several elements that can almost universally increase the chances for successful teacher development and create a powerful and positive school community. The following three sections detail the range of best practices found by researchers to be critical for ensuring educator growth and success:

Ask Edutopia AI BETA

Effective administrator and teacher leadership.

Leadership is second only to teaching among school-related factors that can improve student achievement, and it tends to show greatest impact in traditionally underserved schools ( Leithwood, Seashore Louis, Anderson, and Wahlstrom, 2004 ). Superintendents, principals, and others in positions of authority in school systems are instrumental in providing the vision, time, and resources to support continual professional learning, a positive school climate, and success for all students (Leithwood et al., 2004; The Wallace Foundation, 2012 ). Research shows that the following features of effective leadership can improve student achievement (Leithwood et al., 2004; Vescio, Ross, and Adams, 2008 ; The Wallace Foundation, 2012):

Great leaders focus on developing people's capacities rather than their limitations (Leithwood et al., 2004; Alliance for Excellent Education, 2011 ). Schools that foster trust among parents, teachers, and school leaders are more likely to see academic improvement than schools that do little or fail to foster trust ( Bryk and Schneider, 2003 ).

Teacher leadership is also critical for school improvement efforts to succeed. Accomplished teachers are most knowledgeable about how students in their school or district learn, and thus they are ideal candidates to lead professional-learning and curriculum development efforts (Vescio et al., 2008; Webster-Wright, 2009 ; Accomplished California Teachers, 2012 ). Teacher-advancement systems that effectively identify and support quality teaching include the following features (Accomplished California Teachers, 2012; Darling-Hammond, 2012):

To promote student learning and achievement, research indicates that teacher advancement systems should compensate teachers for their expert contributions, particularly in economically disadvantaged schools where teaching challenges tend to be greater (Accomplished California Teachers, 2012). Finally, researchers discourage the use of value-added modeling in teacher evaluation practices due to their low levels of statistical reliability across years and limited validity for detecting individual teacher effects (Darling-Hammond, 2012).

Job-Embedded Professional Development

When teachers receive well-designed professional development, an average of 49 hours spread over six to 12 months, they can increase student achievement by as much as 21 percentile points ( Yoon, Duncan, Lee, Scarloss, and Shapley, 2007 ). On the other hand, one-shot, "drive-by," or fragmented, "spray-and-pray" workshops lasting 14 hours or less show no statistically significant effect on student learning ( Darling-Hammond, Wei, Andree, Richardson, and Orphanos, 2009 ). Above all, it is most important to remember that effective professional-development programs are job-embedded and provide teachers with five critical elements (Darling-Hammond et al., 2009):

Research on professional development for teachers has shifted in the last decade from delivering and evaluating professional-development programs to focusing more on authentic teacher learning and the conditions that support it (Webster-Wright, 2009). In the next section, we discuss models of professional learning that focus on supporting continual professional learning and community-based feedback cycles that help teachers to critically and collaboratively examine and refine their practices.

Professional Learning Communities

Professional learning communities (PLCs) or networks (PLNs) are groups of teachers that share and critically interrogate their practices in an ongoing, reflective, collaborative, inclusive, learning-oriented, and growth-promoting way to mutually enhance teacher and student learning ( Stoll, Bolam, McMahon, Wallace, and Thomas, 2006 ). PLCs go a step beyond professional development by providing teachers with not just skills and knowledge to improve their teaching practices but also an ongoing community that values each teacher's experiences in their own classrooms and uses those experiences to guide teaching practices and improve student learning (Vescio et al., 2008). Research shows that when professional learning communities demonstrate four key characteristics, they can improve teaching practice and student achievement in reading, writing, math, science, and social studies subject tests (Vescio et al., 2008):

In the following sections, we discuss several practices of professional learning communities that have received consistent support:

Video-based reflections: Using video to reflect upon teaching practice has been shown by several studies to improve teaching practice or student achievement ( Allen, Pianta, Gregory, Mikami, and Lun, 2011 ; Brantlinger, Sherin, and Linsenmeier, 2011 ; Roth, Garnier, Chen, Lemmens, Schwille, and Wickler, 2011 ). In one case study, teachers met regularly to develop video clips of their best teaching practices for the National Board Certification application (Brantlinger et al., 2011). This resulted in the teachers engaging in intensive discussions about mathematical discourse while collaboratively and substantively examining each other's practices (Brantlinger et al., 2011). Similarly, in a case study of four middle school math teachers who participated in a yearlong series of ten video club meetings to reflect on their classrooms, teachers in the video club "came to use video not as a resource for evaluating each other's practices, but rather as a resource for trying to better understand the process of teaching and learning" in a supportive, nonthreatening setting ( Sherin and Han, 2004 ). MyTeachingPartner-Secondary (MTP-S) is a coaching system that provides a library of videos showing effective teaching, as well as personalized Web-based feedback videos of teaching practice using the research-based CLASS-S scoring system to define effective student-teacher interactions (Allen et al., 2011). In a randomized controlled experiment of 78 secondary school teachers and 2,237 students, MTP-S improved teacher-student interactions and increased students' performance on standardized tests by nine percentile points (Allen et al., 2011). Science Teachers Learning through Lesson Analysis (STeLLA) is a professional-development program for upper-elementary school science teachers in which teachers develop two lenses for analyzing teaching, the "Student Thinking Lens" and the "Science Content Storyline Lens," to analyze videos of teaching practice. In an experiment with 48 teachers and 1,490 upper-elementary students, STeLLA improved science teaching and science content knowledge among students and teachers (Roth et al., 2011).

Lesson study: Lesson study is a form of Japanese professional development that engages teachers in collaborative analysis of lessons. It has grown rapidly in the United States since being introduced in 1999 ( Lewis, Perry, and Murata, 2006 ). One purpose of lesson study is to continually improve the experiences that teachers provide for their students. Teachers come together to work on three main activities: (1) identifying a lesson study goal, (2) conducting a small number of study lessons that explore this goal, and (3) reflecting about the process (including producing written reports). In one California school district, lesson study began when an instructional improvement coordinator and a math coach sent an open letter inviting teachers to participate in lesson study during the 2000-01 school year. In the first year, 26 teachers responded, and six years later, the school was still continuing the program. Student achievement data at Highlands Elementary School suggest that lesson study is paying off for students ( Lewis, Perry, Hurd, and O'Connell, 2006 ). Lesson study is used in the majority of elementary schools and middle schools in Japan but is rare in high schools ( Yoshida, 2002 ). For materials to start a lesson study community, check out these resources by Makoto Yoshida, whose 1999 dissertation brought the practice to the attention of U.S. educators, and Catherine Lewis, who conducts academic research on lesson study .

Mentoring programs: A body of research indicates that mentoring programs can increase teacher retention, satisfaction, and student achievement ( Ingersoll and Strong, 2011 ), as well as reduce feelings of isolation, particularly for early-career teachers ( Beltman, Mansfield, and Price, 2011 ). For example, a quasi-experimental study by the Educational Testing Service found that teachers with a high level of engagement in a large-scale mentoring program (California Formative Assessment and Support System for Teachers) improved both teaching practices and student achievement, producing an effect size equivalent to half a year's growth ( Thompson, Goe, Paek, and Ponte, 2004 ). Mentor relationships are most successful when the mentor is positive, pro-social, professional, and from the same teaching area (Beltman et al., 2011).

Grade-level teams: Grade-level teams focused on student learning have also been supported by research. In a quasi-experimental study in nine Title I schools, principals and teacher leaders used explicit protocols for leading grade-level learning teams, resulting in students outperforming their peers in six matched schools on standardized achievement tests ( Gallimore, Ermeling, Saunders, and Goldenberg, 2009 ). These outcomes were more likely for teams led by a trained peer-facilitator, teaching similar content, in stable settings in which to engage in ongoing improvement, and using an inquiry-focused protocol (such as identifying student needs, formulating instructional plans, and using evidence to refine instruction) (Gallimore et al., 2009).

Teacher Development Research Table of Contents:

• Introduction
• Keys to Educator Success
• Avoiding Pitfalls
• Annotated Bibliography

What Works—and What Doesn’t—in Teacher PD

• Share article

When done right, professional development can improve teacher practice and student experiences. But when done wrong, it can have little to no impact and end up frustrating teachers who don’t see any relevance to their work. And it’s all part of a costly, $18 billion market with little quality control.

A new paper, published by the Research Partnership for Professional Learning and written by researchers at Harvard Graduate School of Education and Brown University, examines the literature to understand what works in the field of professional development—and, just as importantly, what doesn’t.

“Teachers in different schools, in different subject areas, in different districts have very different experiences with their professional learning,” said John Papay, an associate professor of education and economics at Brown and a co-author of the paper. “Some of it, we know, can be effective, and some of it, we know, isn’t effective. The challenge is, how do we maintain this investment in and emphasis on professional learning and teacher development throughout the career while also working to make it more effective?”

The research review analyzed both individual studies and syntheses of teacher professional learning, relying mostly on studies that identified a causal impact of the PD on teaching and learning. However, the researchers noted that their review cannot say with certainty that the PD formats and contents are the sole factors behind any success with student outcomes.

It finds that the most effective forms of professional development focus on improving what teachers do in classrooms—their day-to-day practice. It also has an element of accountability involved, so teachers are motivated to change and improve.

Here are five takeaways from the report.

1. PD should focus on instructional practices rather than content knowledge

Over the past two decades, professional development has focused on building teachers’ content knowledge, said Heather Hill, a principal investigator and professor at Harvard Graduate School of Education and a co-author of the paper. The idea was that if teachers have a firm understanding of, say, how fractions work, they will be better at teaching fractions.

But the body of literature suggests that’s not necessarily the case, Hill said, adding that the realization was “personally a little earth-shattering.”

Instead, professional development that focused on changing teachers’ instructional practice—such as by identifying key teaching strategies and providing support for carrying out those changes in the classroom—was found to be more effective for improving student outcomes.

One study in the review directly compared elementary science PD that focused on deepening content knowledge to PD that was focused on analyzing videos of lessons. Teachers spent the same amount of time in both professional development experiences. Students of teachers who did the lesson-analysis PD outperformed students of teachers who did the content-deepening PD by 20 percentile points on a research-developed assessment.

The researchers hypothesized that content-focused professional development might not last long enough for teachers to learn enough about the subject to truly make a difference in their instruction. Also, those types of PD programs often don’t offer much support for the day-to-day practice, and teachers need to be able to connect their learning to their existing curriculum materials or lesson plans, the researchers said.

2. PD should prioritize concrete materials for practice over general principles

There are two approaches toward PD that can be at odds. The first is to give teachers materials like curricula, lessons, and assessment items that offer concrete ways to reach the goal, but may leave them without a strong understanding of the learning philosophy behind the new approach. The second is to emphasize more general principals to promote broader and more lasting changes in instruction, but leave it up to the teachers themselves to integrate those changes in their existing lessons, materials, and assessments.

For example, one approach to PD could focus on helping teachers learn how to use formative assessment items in their classroom and giving them some models; the other approach might emphasize design principles so teachers can create their own new formative assessment items.

The research review found that focusing PD on concrete materials is more effective than teaching general principles, which usually ends up requiring teachers to do additional work on their own time. PD that provides support for the day-to-day is more likely to increase uptake and improve the quality of the implementation.

“It needs to be job-relevant in a way that teachers can see how it improves their practice and is not asking them to do extra work,” Hill said.

3. Have follow-up meetings after PD or coaching

A low-cost way to boost the effectiveness of a PD program is to add a post-implementation follow-up meeting, the research review found. Teachers can share their experiences implementing the practices learned and receive feedback from colleagues and program facilitators. They can also ask questions and voice concerns about parts of the new program that are particularly challenging to implement.

These sessions are typically collaborative, so teachers can share ideas with one another and perhaps even improve the program by customizing it to meet the needs of their students and school.

Also, the paper notes, follow-up sessions offer some accountability—teachers are more likely to implement the practices if they know they will need to report on how it went to their colleagues and facilitators.

4. PD should help teachers build relationships with students

Past research has shown that strong teacher-student relationships can lead to higher student academic engagement, better attendance, better grades, fewer disruptive behaviors and suspensions, and lower school dropout rates. Those effects were strong even after controlling for differences in students’ individual, family, and school backgrounds.

These relationships can be fostered and improved through targeted professional learning, the researchers found.

The University of Virginia’s school of education offers professional-development support focused on improving teacher-student interactions. The program, called MyTeachingPartner , has been associated with student gains in learning and the closing of the racial discipline gap in high school.

Hill said she has witnessed facilitators in those trainings share easy-to-implement strategies for teachers to better connect with students. For example, a facilitator urged teachers to stand at the door as students file in at the start of the class, greeting them individually and asking questions about their life outside of school (like how a basketball game went).

These are “on its face, very simple strategies that actually can be pretty powerful,” Hill said.

5. Coaching and teacher collaboration are key strategies

The research review emphasized the effectiveness of both peer collaboration and coaching. Evidence suggests that teachers can and do learn from each other, and that when schools promote collaboration, teacher practice and student outcomes improve. Coaching—which can include modeling instruction, co-planning lessons, direct feedback, and other consultations and support—has also been found to successfully improve classroom instructional quality and student outcomes.

However, the design of these practices matters. Collaboration should be focused on shared and specific goals for improvement rather than meeting to vaguely improve practice. And teachers should have dedicated and protected time to work and learn together.

Meanwhile, coaching is most effective when it’s more focused—when the coaches can focus on working with teachers instead of administrative duties, and when the coaches also receive some professional development and leadership support.

Yet the realities of school operations these days often don’t allow for these conditions, the researchers said. Many schools are struggling to staff classrooms, and coaches are often tapped to act as substitute teachers, pulling them away from the core functions of their jobs. And collaboration time can be put at risk when teachers have to cover other classrooms.

“There’s coaching as it is in the literature, and coaching that exists in schools,” Hill said.

Educators have a lot on their plates this school year, and teacher stress levels are still high, surveys show. Still, teachers are tasked with helping students recover unfinished learning as a result of the pandemic , making effective professional development more important than ever, the researchers said.

“Finding opportunities for teachers to engage in professional learning seems particularly critical now because that type of support, that type of ongoing development ... leads to teachers feeling more satisfied [in their jobs, which can] alleviate burnout,” Papay said. “Cutting out professional learning or not prioritizing it will, in some ways, lead to larger challenges downstream.”

Sign Up for The Savvy Principal

Edweek top school jobs.

Sign Up & Sign In

• Search Menu
• Sign in through your institution
• Browse content in A - General Economics and Teaching
• Browse content in A1 - General Economics
• A12 - Relation of Economics to Other Disciplines
• A14 - Sociology of Economics
• Browse content in B - History of Economic Thought, Methodology, and Heterodox Approaches
• Browse content in B4 - Economic Methodology
• B41 - Economic Methodology
• Browse content in C - Mathematical and Quantitative Methods
• Browse content in C1 - Econometric and Statistical Methods and Methodology: General
• C18 - Methodological Issues: General
• Browse content in C2 - Single Equation Models; Single Variables
• C21 - Cross-Sectional Models; Spatial Models; Treatment Effect Models; Quantile Regressions
• Browse content in C3 - Multiple or Simultaneous Equation Models; Multiple Variables
• C38 - Classification Methods; Cluster Analysis; Principal Components; Factor Models
• Browse content in C5 - Econometric Modeling
• C59 - Other
• Browse content in C8 - Data Collection and Data Estimation Methodology; Computer Programs
• C80 - General
• C81 - Methodology for Collecting, Estimating, and Organizing Microeconomic Data; Data Access
• C83 - Survey Methods; Sampling Methods
• Browse content in C9 - Design of Experiments
• C93 - Field Experiments
• Browse content in D - Microeconomics
• Browse content in D0 - General
• D02 - Institutions: Design, Formation, Operations, and Impact
• D03 - Behavioral Microeconomics: Underlying Principles
• D04 - Microeconomic Policy: Formulation; Implementation, and Evaluation
• Browse content in D1 - Household Behavior and Family Economics
• D10 - General
• D12 - Consumer Economics: Empirical Analysis
• D14 - Household Saving; Personal Finance
• Browse content in D2 - Production and Organizations
• D22 - Firm Behavior: Empirical Analysis
• D24 - Production; Cost; Capital; Capital, Total Factor, and Multifactor Productivity; Capacity
• Browse content in D3 - Distribution
• D31 - Personal Income, Wealth, and Their Distributions
• Browse content in D6 - Welfare Economics
• D61 - Allocative Efficiency; Cost-Benefit Analysis
• D62 - Externalities
• D63 - Equity, Justice, Inequality, and Other Normative Criteria and Measurement
• Browse content in D7 - Analysis of Collective Decision-Making
• D72 - Political Processes: Rent-seeking, Lobbying, Elections, Legislatures, and Voting Behavior
• D73 - Bureaucracy; Administrative Processes in Public Organizations; Corruption
• D74 - Conflict; Conflict Resolution; Alliances; Revolutions
• Browse content in D8 - Information, Knowledge, and Uncertainty
• D83 - Search; Learning; Information and Knowledge; Communication; Belief; Unawareness
• D85 - Network Formation and Analysis: Theory
• D86 - Economics of Contract: Theory
• Browse content in D9 - Micro-Based Behavioral Economics
• D91 - Role and Effects of Psychological, Emotional, Social, and Cognitive Factors on Decision Making
• D92 - Intertemporal Firm Choice, Investment, Capacity, and Financing
• Browse content in E - Macroeconomics and Monetary Economics
• Browse content in E2 - Consumption, Saving, Production, Investment, Labor Markets, and Informal Economy
• E23 - Production
• E24 - Employment; Unemployment; Wages; Intergenerational Income Distribution; Aggregate Human Capital; Aggregate Labor Productivity
• Browse content in E4 - Money and Interest Rates
• E42 - Monetary Systems; Standards; Regimes; Government and the Monetary System; Payment Systems
• Browse content in E5 - Monetary Policy, Central Banking, and the Supply of Money and Credit
• E52 - Monetary Policy
• E58 - Central Banks and Their Policies
• Browse content in E6 - Macroeconomic Policy, Macroeconomic Aspects of Public Finance, and General Outlook
• E60 - General
• E61 - Policy Objectives; Policy Designs and Consistency; Policy Coordination
• E62 - Fiscal Policy
• E65 - Studies of Particular Policy Episodes
• Browse content in F - International Economics
• Browse content in F0 - General
• F01 - Global Outlook
• Browse content in F1 - Trade
• F10 - General
• F11 - Neoclassical Models of Trade
• F13 - Trade Policy; International Trade Organizations
• F14 - Empirical Studies of Trade
• F15 - Economic Integration
• F19 - Other
• Browse content in F2 - International Factor Movements and International Business
• F21 - International Investment; Long-Term Capital Movements
• F22 - International Migration
• F23 - Multinational Firms; International Business
• Browse content in F3 - International Finance
• F32 - Current Account Adjustment; Short-Term Capital Movements
• F34 - International Lending and Debt Problems
• F35 - Foreign Aid
• F36 - Financial Aspects of Economic Integration
• Browse content in F4 - Macroeconomic Aspects of International Trade and Finance
• F41 - Open Economy Macroeconomics
• F42 - International Policy Coordination and Transmission
• F43 - Economic Growth of Open Economies
• Browse content in F5 - International Relations, National Security, and International Political Economy
• F50 - General
• F52 - National Security; Economic Nationalism
• F53 - International Agreements and Observance; International Organizations
• F55 - International Institutional Arrangements
• Browse content in F6 - Economic Impacts of Globalization
• F61 - Microeconomic Impacts
• F63 - Economic Development
• F66 - Labor
• Browse content in G - Financial Economics
• Browse content in G0 - General
• G01 - Financial Crises
• Browse content in G1 - General Financial Markets
• G10 - General
• G15 - International Financial Markets
• G18 - Government Policy and Regulation
• Browse content in G2 - Financial Institutions and Services
• G20 - General
• G21 - Banks; Depository Institutions; Micro Finance Institutions; Mortgages
• G22 - Insurance; Insurance Companies; Actuarial Studies
• G23 - Non-bank Financial Institutions; Financial Instruments; Institutional Investors
• G28 - Government Policy and Regulation
• Browse content in G3 - Corporate Finance and Governance
• G32 - Financing Policy; Financial Risk and Risk Management; Capital and Ownership Structure; Value of Firms; Goodwill
• G33 - Bankruptcy; Liquidation
• G38 - Government Policy and Regulation
• Browse content in H - Public Economics
• Browse content in H1 - Structure and Scope of Government
• H11 - Structure, Scope, and Performance of Government
• Browse content in H2 - Taxation, Subsidies, and Revenue
• H20 - General
• H23 - Externalities; Redistributive Effects; Environmental Taxes and Subsidies
• H25 - Business Taxes and Subsidies
• H26 - Tax Evasion and Avoidance
• H27 - Other Sources of Revenue
• Browse content in H3 - Fiscal Policies and Behavior of Economic Agents
• H31 - Household
• Browse content in H4 - Publicly Provided Goods
• H41 - Public Goods
• H43 - Project Evaluation; Social Discount Rate
• Browse content in H5 - National Government Expenditures and Related Policies
• H52 - Government Expenditures and Education
• H53 - Government Expenditures and Welfare Programs
• H54 - Infrastructures; Other Public Investment and Capital Stock
• H55 - Social Security and Public Pensions
• H56 - National Security and War
• H57 - Procurement
• Browse content in H6 - National Budget, Deficit, and Debt
• H60 - General
• H61 - Budget; Budget Systems
• Browse content in H7 - State and Local Government; Intergovernmental Relations
• H71 - State and Local Taxation, Subsidies, and Revenue
• H75 - State and Local Government: Health; Education; Welfare; Public Pensions
• H77 - Intergovernmental Relations; Federalism; Secession
• Browse content in H8 - Miscellaneous Issues
• H83 - Public Administration; Public Sector Accounting and Audits
• H84 - Disaster Aid
• Browse content in I - Health, Education, and Welfare
• Browse content in I0 - General
• I00 - General
• Browse content in I1 - Health
• I10 - General
• I12 - Health Behavior
• I15 - Health and Economic Development
• I18 - Government Policy; Regulation; Public Health
• Browse content in I2 - Education and Research Institutions
• I20 - General
• I21 - Analysis of Education
• I22 - Educational Finance; Financial Aid
• I24 - Education and Inequality
• I25 - Education and Economic Development
• I28 - Government Policy
• Browse content in I3 - Welfare, Well-Being, and Poverty
• I30 - General
• I31 - General Welfare
• I32 - Measurement and Analysis of Poverty
• I38 - Government Policy; Provision and Effects of Welfare Programs
• Browse content in J - Labor and Demographic Economics
• Browse content in J0 - General
• J01 - Labor Economics: General
• J08 - Labor Economics Policies
• Browse content in J1 - Demographic Economics
• J10 - General
• J11 - Demographic Trends, Macroeconomic Effects, and Forecasts
• J12 - Marriage; Marital Dissolution; Family Structure; Domestic Abuse
• J13 - Fertility; Family Planning; Child Care; Children; Youth
• J15 - Economics of Minorities, Races, Indigenous Peoples, and Immigrants; Non-labor Discrimination
• J16 - Economics of Gender; Non-labor Discrimination
• J17 - Value of Life; Forgone Income
• J18 - Public Policy
• Browse content in J2 - Demand and Supply of Labor
• J21 - Labor Force and Employment, Size, and Structure
• J22 - Time Allocation and Labor Supply
• J23 - Labor Demand
• J24 - Human Capital; Skills; Occupational Choice; Labor Productivity
• J26 - Retirement; Retirement Policies
• J28 - Safety; Job Satisfaction; Related Public Policy
• Browse content in J3 - Wages, Compensation, and Labor Costs
• J38 - Public Policy
• Browse content in J4 - Particular Labor Markets
• J48 - Public Policy
• Browse content in J5 - Labor-Management Relations, Trade Unions, and Collective Bargaining
• J58 - Public Policy
• Browse content in J6 - Mobility, Unemployment, Vacancies, and Immigrant Workers
• J61 - Geographic Labor Mobility; Immigrant Workers
• J62 - Job, Occupational, and Intergenerational Mobility
• J63 - Turnover; Vacancies; Layoffs
• J68 - Public Policy
• Browse content in J8 - Labor Standards: National and International
• J88 - Public Policy
• Browse content in K - Law and Economics
• Browse content in K2 - Regulation and Business Law
• K23 - Regulated Industries and Administrative Law
• Browse content in K3 - Other Substantive Areas of Law
• K32 - Environmental, Health, and Safety Law
• K34 - Tax Law
• Browse content in K4 - Legal Procedure, the Legal System, and Illegal Behavior
• K40 - General
• K41 - Litigation Process
• K42 - Illegal Behavior and the Enforcement of Law
• Browse content in L - Industrial Organization
• Browse content in L1 - Market Structure, Firm Strategy, and Market Performance
• L11 - Production, Pricing, and Market Structure; Size Distribution of Firms
• L14 - Transactional Relationships; Contracts and Reputation; Networks
• L16 - Industrial Organization and Macroeconomics: Industrial Structure and Structural Change; Industrial Price Indices
• Browse content in L2 - Firm Objectives, Organization, and Behavior
• L20 - General
• L23 - Organization of Production
• L25 - Firm Performance: Size, Diversification, and Scope
• L26 - Entrepreneurship
• Browse content in L3 - Nonprofit Organizations and Public Enterprise
• L33 - Comparison of Public and Private Enterprises and Nonprofit Institutions; Privatization; Contracting Out
• Browse content in L5 - Regulation and Industrial Policy
• L51 - Economics of Regulation
• L52 - Industrial Policy; Sectoral Planning Methods
• Browse content in L9 - Industry Studies: Transportation and Utilities
• L94 - Electric Utilities
• L97 - Utilities: General
• L98 - Government Policy
• Browse content in M - Business Administration and Business Economics; Marketing; Accounting; Personnel Economics
• Browse content in M5 - Personnel Economics
• M53 - Training
• Browse content in N - Economic History
• Browse content in N1 - Macroeconomics and Monetary Economics; Industrial Structure; Growth; Fluctuations
• N10 - General, International, or Comparative
• Browse content in N3 - Labor and Consumers, Demography, Education, Health, Welfare, Income, Wealth, Religion, and Philanthropy
• N35 - Asia including Middle East
• Browse content in N5 - Agriculture, Natural Resources, Environment, and Extractive Industries
• N55 - Asia including Middle East
• N57 - Africa; Oceania
• Browse content in N7 - Transport, Trade, Energy, Technology, and Other Services
• N77 - Africa; Oceania
• Browse content in O - Economic Development, Innovation, Technological Change, and Growth
• Browse content in O1 - Economic Development
• O10 - General
• O11 - Macroeconomic Analyses of Economic Development
• O12 - Microeconomic Analyses of Economic Development
• O13 - Agriculture; Natural Resources; Energy; Environment; Other Primary Products
• O14 - Industrialization; Manufacturing and Service Industries; Choice of Technology
• O15 - Human Resources; Human Development; Income Distribution; Migration
• O16 - Financial Markets; Saving and Capital Investment; Corporate Finance and Governance
• O17 - Formal and Informal Sectors; Shadow Economy; Institutional Arrangements
• O18 - Urban, Rural, Regional, and Transportation Analysis; Housing; Infrastructure
• O19 - International Linkages to Development; Role of International Organizations
• Browse content in O2 - Development Planning and Policy
• O20 - General
• O22 - Project Analysis
• O23 - Fiscal and Monetary Policy in Development
• O24 - Trade Policy; Factor Movement Policy; Foreign Exchange Policy
• O25 - Industrial Policy
• Browse content in O3 - Innovation; Research and Development; Technological Change; Intellectual Property Rights
• O31 - Innovation and Invention: Processes and Incentives
• O32 - Management of Technological Innovation and R&D
• O33 - Technological Change: Choices and Consequences; Diffusion Processes
• O34 - Intellectual Property and Intellectual Capital
• O38 - Government Policy
• Browse content in O4 - Economic Growth and Aggregate Productivity
• O40 - General
• O41 - One, Two, and Multisector Growth Models
• O43 - Institutions and Growth
• O44 - Environment and Growth
• O47 - Empirical Studies of Economic Growth; Aggregate Productivity; Cross-Country Output Convergence
• Browse content in O5 - Economywide Country Studies
• O55 - Africa
• O57 - Comparative Studies of Countries
• Browse content in P - Economic Systems
• Browse content in P1 - Capitalist Systems
• P14 - Property Rights
• Browse content in P2 - Socialist Systems and Transitional Economies
• P26 - Political Economy; Property Rights
• Browse content in P3 - Socialist Institutions and Their Transitions
• P30 - General
• Browse content in P4 - Other Economic Systems
• P43 - Public Economics; Financial Economics
• P46 - Consumer Economics; Health; Education and Training; Welfare, Income, Wealth, and Poverty
• P48 - Political Economy; Legal Institutions; Property Rights; Natural Resources; Energy; Environment; Regional Studies
• Browse content in Q - Agricultural and Natural Resource Economics; Environmental and Ecological Economics
• Browse content in Q0 - General
• Q01 - Sustainable Development
• Browse content in Q1 - Agriculture
• Q10 - General
• Q12 - Micro Analysis of Farm Firms, Farm Households, and Farm Input Markets
• Q13 - Agricultural Markets and Marketing; Cooperatives; Agribusiness
• Q14 - Agricultural Finance
• Q15 - Land Ownership and Tenure; Land Reform; Land Use; Irrigation; Agriculture and Environment
• Q16 - R&D; Agricultural Technology; Biofuels; Agricultural Extension Services
• Q17 - Agriculture in International Trade
• Q18 - Agricultural Policy; Food Policy
• Browse content in Q2 - Renewable Resources and Conservation
• Q25 - Water
• Browse content in Q3 - Nonrenewable Resources and Conservation
• Q30 - General
• Q33 - Resource Booms
• Browse content in Q4 - Energy
• Q41 - Demand and Supply; Prices
• Q43 - Energy and the Macroeconomy
• Browse content in Q5 - Environmental Economics
• Q51 - Valuation of Environmental Effects
• Q52 - Pollution Control Adoption Costs; Distributional Effects; Employment Effects
• Q54 - Climate; Natural Disasters; Global Warming
• Q56 - Environment and Development; Environment and Trade; Sustainability; Environmental Accounts and Accounting; Environmental Equity; Population Growth
• Q57 - Ecological Economics: Ecosystem Services; Biodiversity Conservation; Bioeconomics; Industrial Ecology
• Q58 - Government Policy
• Browse content in R - Urban, Rural, Regional, Real Estate, and Transportation Economics
• Browse content in R1 - General Regional Economics
• R11 - Regional Economic Activity: Growth, Development, Environmental Issues, and Changes
• R12 - Size and Spatial Distributions of Regional Economic Activity
• R13 - General Equilibrium and Welfare Economic Analysis of Regional Economies
• R14 - Land Use Patterns
• Browse content in R2 - Household Analysis
• R20 - General
• R23 - Regional Migration; Regional Labor Markets; Population; Neighborhood Characteristics
• R28 - Government Policy
• Browse content in R3 - Real Estate Markets, Spatial Production Analysis, and Firm Location
• R38 - Government Policy
• Browse content in R4 - Transportation Economics
• R40 - General
• R41 - Transportation: Demand, Supply, and Congestion; Travel Time; Safety and Accidents; Transportation Noise
• R48 - Government Pricing and Policy
• Browse content in R5 - Regional Government Analysis
• R52 - Land Use and Other Regulations
• Browse content in Y - Miscellaneous Categories
• Y8 - Related Disciplines
• Browse content in Z - Other Special Topics
• Browse content in Z1 - Cultural Economics; Economic Sociology; Economic Anthropology
• Z13 - Economic Sociology; Economic Anthropology; Social and Economic Stratification
• Advance articles
• Author Guidelines
• Open Access
• About The World Bank Research Observer
• About the World Bank
• Editorial Board
• Advertising and Corporate Services
• Self-Archiving Policy
• Dispatch Dates
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

• < Previous

Teacher Professional Development around the World: The Gap between Evidence and Practice

• Article contents
• Figures & tables
• Supplementary Data

Anna Popova, David K Evans, Mary E Breeding, Violeta Arancibia, Teacher Professional Development around the World: The Gap between Evidence and Practice, The World Bank Research Observer , Volume 37, Issue 1, February 2022, Pages 107–136, https://doi.org/10.1093/wbro/lkab006

• Permissions Icon Permissions

Many teachers in low- and middle-income countries lack the skills to teach effectively, and professional development (PD) programs are the principal tool that governments use to upgrade those skills. At the same time, few PD programs are evaluated, and those that are evaluated show highly varying results. This paper proposes a set of indicators—the In-Service Teacher Training Survey Instrument—to standardize reporting on teacher PD programs. An application of the instrument to 33 rigorously evaluated PD programs shows that programs that link participation to career incentives, have a specific subject focus, incorporate lesson enactment in the training, and include initial face-to-face training tend to show higher student learning gains. In qualitative interviews, program implementers also report follow-up visits as among the most effective characteristics of their professional development programs. This paper then uses the instrument to present novel data on a sample of 139 government-funded, at-scale professional development programs across 14 countries. The attributes of most at-scale teacher professional development programs differ sharply from those of programs that evidence suggests are effective, with fewer incentives to participate in PD, fewer opportunities to practice new skills, and less follow-up once teachers return to their classrooms.

Good teachers have a major impact on student performance, both over the course of the school year ( Araujo et al. 2016 ) and into adulthood ( Chetty, Friedman, and Rockoff 2014 ). However, teachers in low- and middle-income countries often lack the skills they need to teach students effectively. Across seven African countries, only seven percent of fourth-grade teachers had the minimum knowledge necessary to teach language; in four countries, the statistic was zero percent. For math teaching, 68 percent had the minimum knowledge needed to teach math—higher than the seven percent for language, but still leaving one in three teachers with insufficient knowledge. Teachers also scored woefully low in terms of pedagogical knowledge—their ability to prepare a lesson, formulate questions that would elicit student knowledge effectively, and their performance in the classroom ( Bold et al. 2017 ).

The principal tool that countries across the income spectrum use to improve the knowledge and skills of their practicing teachers is professional development (PD), which refers to on-the-job training activities ranging from formal, lecture-style training to mentoring and coaching. However, few PD programs are rigorously evaluated, and among those that are, the evidence of their effectiveness is wildly mixed. Some programs are effective: training teachers to provide literacy instruction using students’ mother tongue in Uganda and training teachers to evaluate student performance more regularly and adjust teaching based on those evaluations in Liberia both had sizeable impacts on student reading ability ( Piper and Korda 2011 ; Kerwin and Thornton 2021 ). Others demonstrate opposite results: a large-scale, government-implemented PD program in China had zero impact on teacher knowledge, teaching practices, or student learning outcomes ( Loyalka et al. 2019 ), and a program that trained teachers to engage their middle school math students more actively in learning in Costa Rica resulted in worse learning outcomes for students ( Berlinski and Busso 2017 ). Indeed, there is much more variation in effectiveness across teacher training programs than across education programs more broadly ( McEwan 2015 ; Evans and Popova 2016a ). With this limited and highly variable evidence, policymakers and practitioners may be left puzzled as to how to structure teacher PD programs effectively.

In this paper, we propose a set of indicators—the In-service Teacher Training Survey Instrument, or ITTSI—to allow comparisons across teacher PD programs with varying impacts. On average, existing studies of PD programs only report on about half of these indicators. We supplement that information through interviews with implementors of evaluated PD programs. We compare the characteristics of 33 rigorously evaluated PD programs to identify which characteristics are associated with larger student learning gains. We then gather data from 139 government-funded, at-scale PD programs across 14 countries. Like most at-scale government programs, none of these programs have been evaluated rigorously. We compare the two samples to examine whether the PD programs that most teachers actually experience exhibit similar characteristics to those of PD programs that have been evaluated and shown to produce sizeable student learning gains.

When we apply our instrument to evaluated PD programs, results suggest that programs deliver high student learning gains when they link participation in PD to incentives such as promotion or salary implications, when they have a specific subject focus, when teachers practice enacting lessons during the training, and when training has at least an initial face-to-face aspect. Meanwhile, program implementers highlight two characteristics of effective training in interviews: mentoring follow-up visits after the PD training, and complementary materials such as structured lesson plans to help teachers apply what they have learned during PD.

When we subsequently use the ITTSI to characterize a sample of at-scale, government-funded PD programs around the world, we find a divergence in the characteristics common to these programs and those that typify evaluated programs that were found to be effective. Relative to top-performing PD programs—defined as those found to be the most effective at increasing student learning—very few at-scale PD programs are linked to any sort of career opportunities, such as promotion or salary implications. Similarly, in-school follow-up support and including time to practice with other teachers is less common among at-scale PD programs. This highlights a substantial gap between the kind of teacher PD supported by research and that currently being provided by many government-funded, at-scale programs.

These results have implications for both researchers and policymakers. For researchers, future evaluations will contribute much more to an understanding of how to improve teachers’ skills if they report more details of the characteristics of the PD programs. Our proposed set of indicators can serve as a guide. For policymakers, at-scale PD programs should incorporate more aspects of successful, evaluated PD programs, such as incentives, practice, and follow-up in-school support. For both, more programs can be evaluated at scale, using government delivery systems, in order to improve the skills of teachers in the future.

Conceptual Framework

The defining attributes of teacher professional development programs fall principally into three categories. The first is the content of the PD program: What is taught? The second is the delivery of the PD program: Who is teaching, when, and for how long? The third is the organization of the program beyond content and delivery: What are the scale and resources of the program? Are there incentives for participation? Was it designed based on a diagnostic of teachers? In this section, we discuss the theory behind each of these three categories.

On the content, PD programs focused on subject-specific pedagogy are likely to be most effective. General pedagogical knowledge—i.e., broad strategies of classroom management and organization—may contribute to student learning, driving the recent development of a range of classroom observation instruments ( La Paro and Pianta 2003 ; Molina et al. 2018 ). However, different subjects require radically different pedagogies ( Shulman 1986 ; Villegas-Reimers 2003 ). A highly scripted approach may work to teach early grade reading, whereas teaching science or civics in later grades—for example—may require more flexible approaches. PD programs that focus on arming teachers with subject-specific pedagogy are thus likely to make the largest contribution to student learning.

With respect to the delivery, the method, trainers, duration, and location of instruction all play a role. First, because working, professional teachers are the students in PD, principles of adult education are relevant to the method of instruction. Adult education tends to work best with clear applications rather than a theoretical focus ( Cardemil 2001 ; Knowles, Holton, and Swanson 2005 ). The method of instruction should include concrete, realistic goals ( Baker and Smith 1999 ) and the teaching of formative evaluation so that teachers can effectively evaluate their own progress towards their teaching goals ( Bourgeois and Nizet 1997 ). Second, the quality of trainers—i.e., those providing the PD—is crucial to learning ( Knowles, Holton, and Swanson 2005 ). In terms of the delivery of PD, this calls into question the common cascade model of PD in low-income environments, in which both information and pedagogical ability may be diluted as a master trainer trains another individual as a trainer, who may go on to train another trainer below her, and so forth.

Third, on the duration of instruction, there is no theoretical consensus on exactly how long training should last, although there is suggestive empirical evidence in the literature in favor of sustained contact over a significant period of time and against brief, one-time workshops ( Desimone 2009 ). Fourth, on the location of instruction, teacher PD in the school (“embedded”) is likely to be most effective so that participating teachers can raise concrete problems that they face in the local environment, and they can also receive feedback on actual teaching ( Wood and McQuarrie 1999 ). However, this will depend on the environment. In very difficult teaching environments, some degree of training outside the school may facilitate focus on the part of the trainees ( Kraft and Papay 2014 ).

Finally, the organization of the PD—which includes overarching aspects such as who is organizing it, for whom, and how—provides an important backdrop when we consider any PD program. This includes aspects such as the scale, cost, and targeting of the program. In general, it is predictably easier to provide high-quality PD through smaller scale, higher cost programs that provide more tailored attention to a given teacher. In terms of targeting, teacher PD will work best if it adjusts at different points in the teachers’ careers: One would not effectively teach a brand-new teacher in the same way as one would train a teacher with 20 years of experience ( Huberman 1989 ). Teachers see their greatest natural improvements in the first five years of teaching, which may be an indicator of greater skill plasticity, so there may be benefits to leveraging that time ( TNTP 2015 ).

What Works in High-Income Countries?

A full review of the literature in high-income countries is beyond the scope of this study. However, it may be useful to highlight recent work on in-service teacher PD from the United States—which spends almost $18,000 per teacher and 19 days of teacher time on training each year ( TNTP 2015 )—and other high-income countries, in order to ensure that low- and middle-income countries are not ignoring well-established evidence. Several promising themes that emerge from this work are the importance of making PD specific and practical, providing sustained follow-up support for teachers, and embedding it in the curriculum.

Specific and practical teacher PD finds support from multiple reviews of teacher PD studies in high-income countries, which conclude that concrete, classroom-based programs make the most difference to teachers ( Darling-Hammond et al. 2009 ; Walter and Briggs 2012 ). More recently, a meta-analysis of 196 randomized evaluations of education interventions—not just PD—in the United States that measure student test scores as an outcome examined the impact of both “general” and “managed” professional development, relative to other interventions ( Fryer 2017 ). General PD may focus on classroom management or increasing the rigor of teachers’ knowledge, whereas managed professional development prescribes a specific method, with detailed instructions on implementation and follow-up support. On average, managed PD increased student test scores by 2.5 times (0.052 standard deviations) as much as general PD and was at least as effective as the combined average of all school-based interventions. A recent review of nearly 2,000 impact estimates from 747 randomized controlled trials of education interventions in the United States proposes that an effect size of 0.05 be considered a “medium” effect size, higher than the average effect size, weighted by study sample size ( Kraft 2020 ), which suggests that these are not trivial impacts.

The importance of sustained follow-up support is echoed by another U.S.-focused review, which found that PD programs with significant contact hours (between 30 and 100 in total) over the course of six to twelve months were more effective at raising student test scores ( Yoon et al. 2007 ). Likewise, a narrative review of U.S. studies concluded that the most effective programs are not “one-shot workshops”: they are sustained, intense, and embedded in the curriculum ( Darling-Hammond et al. 2009 ).

Despite these conclusions, the experimental or quasi-experimental evidence is thin, even in high-income countries. The meta-analysis of 196 evaluations of education interventions included just nine PD studies ( Fryer 2017 ), and another review of 1,300 PD studies identified just nine that had pre- and post-test data and some sort of control group ( Yoon et al. 2007 ). Similarly, a review of PD in mathematics found more than 600 studies of math PD interventions, but only 32 used any research design to measure effectiveness, and only five of those were high-quality randomized trials ( Gersten et al. 2014 ). The question of what drives effective teacher PD remains understudied, even in high-income environments.

We expect teachers in lower and middle-income countries to learn in fundamentally similar ways to their high-income counterparts. However, lower resource contexts are typically characterized by more binding cost constraints and lower teacher and coach pedagogical capacity. These challenges may make certain elements of PD programs more and less relevant in lower-income contexts. Teachers and coaches in low- and middle-income countries may benefit from more prescriptive instructions on implementation and, while they too require ongoing follow-up as part of PD, this may need to be provided in lower-cost forms, whether in group sessions, using technology for remote coaching, or training school principals and experienced peer teachers as coaches.

To understand which characteristics of PD programs are associated with student test score gains, and to analyze the degree to which these effective characteristics are incorporated into at-scale PD programs in practice, we first developed a standardized instrument to characterize in-service teacher training. Second, we applied this instrument to already evaluated PD programs to understand which PD characteristics are associated with student learning gains. Third, we applied the survey instrument to a sample of at-scale PD programs to see how these programs line up with what the evidence suggests works in teacher training. The information we present thus comes from two different samples of PD programs: One sample of evaluated PD programs, those with impact evaluations that include student assessment results; and one sample of at-scale , government-funded PD programs. 1 The remainder of this section introduces the instrument briefly before describing its application to each of the two samples.

The In-Service Teacher Training Survey Instrument (ITTSI)

The ITTSI was designed based on the conceptual framework and empirical literature characterized in the previous sections, as well as on the authors’ prior experience studying in-service teacher PD. We drafted an initial list of 51 key indicators to capture details about a range of program characteristics falling into three main categories: Organization, Content, and Delivery, paralleling the three elements of our conceptual framework ( fig. 1 ). We supplement those categories with a fourth category, Perceptions, which we added to collect qualitative data from program implementors.

Summary of the In-Service Teacher Training Survey Instrument (ITTSI)

Source : Authors’ summary of the elements of the In-Service Teacher Training Survey Instrument, as detailed in supplementary online appendices A1 and A2 .

Taking each of these in turn, the Organization section includes items such as the type of organization responsible for the design and implementation of a given teacher training program, to whom the program is targeted, what (if any) complementary materials it provides, the scale of the program, and its cost. The Content section includes indicators capturing the type of knowledge or skills that a given program aims to build among beneficiary teachers, such as whether the program focuses on subject content (and if so, which subject), pedagogy, new technology, classroom management, counseling, assessment, or some combination.

Delivery focuses on indicators capturing program implementation details, such as whether it is delivered through a cascade model, the profile of the trainers who directly train the teachers, the location of the training, the size of the sessions, and the time division between lectures, practice, and other activities. Finally, the Perceptions section includes indicators capturing program implementers’ own perceptions of which elements were responsible for any positive impacts and which were popular or unpopular among teachers. We piloted the draft instrument by using it to collect data on a sample of evaluated programs, and validated its ability to accurately characterize the details of PD programs by sharing our results with a series of expert researchers and practitioners in teacher PD. We updated the indicators in light of this feedback, resulting in a final version of the instrument, which includes 70 indicators plus three pieces of metadata. Further information on the instrument can be found in the supplementary online appendices: Appendix A1 provides a more detailed description of instrument development; appendix A2 presents the final instrument (ITTSI); and appendix A3 presents the Brief In-Service Teacher Training Instrument (BITTSI), a supplementary instrument we developed containing a subset of the 13 most critical questions from the ITTSI based on our reading of the literature.

The ITTSI does not collect extensive data about the broader educational context. Context includes teacher policies (e.g., pre-service training and the structure of the teacher career), other education policies, and the current state of education (e.g., learning and absenteeism rates). Context matters for the impact of teacher PD programs. As a simple example, in a setting where student absenteeism is extremely high, teacher PD programs may have a limited impact on student learning due to few hours of contact between teachers and students. That said, certain principles of teacher PD may translate across cultures, even if the applications vary. Professionals need practice to master skills across contexts, so giving teachers the opportunity to practice lessons during training may be valuable across contexts, even if how they do that may vary. Other survey instruments have been developed and tested broadly to gather a wide range of data on the education system, notably the World Bank's Systems Approach for Better Education Results (SABER) ( Rogers and Demas 2013 ). For a rich view of teacher PD in context, the ITTSI could be complemented with the SABER instrument or other data about the education system.

Applying the ITTSI to Evaluated PD Programs

We searched the existing literature on in-service teacher PD in low- and middle-income countries to identify a sample of PD programs that had been evaluated for their impact on student learning. Our inclusion criteria for the search were impact evaluations of primary and secondary education interventions in low- and middle-income countries that (a) focused primarily on in-service teacher PD or included this as a major component of a broader program, and (b) reported impacts of the program on student test scores in math, language, or science. We included both published and unpublished papers and did not restrict by year of authorship.

In order to identify papers fulfilling the above criteria, we searched a range of databases in 2016 . 2 The search yielded 6,049 results and automatically refined the results by removing exact duplicates from the original results, which reduced the number of results to 4,294. To this we added 20 impact evaluations which mention teacher PD from a recent review ( Evans and Popova 2016a ). We examined the 4,314 results from both sources to exclude articles that—from their title and abstract—were clearly not impact evaluations of teacher training programs. This review process excluded 4,272 results and left 42 full articles to be assessed for eligibility. After going through the full texts, another 18 papers were excluded as the full text revealed that they did not meet the inclusion criteria. This yielded 23 papers, which evaluated 26 different PD programs. In February 2018, we updated this original sample with full articles published between 2016 and 2018 which fit the inclusion criteria. This resulted in seven new papers and teacher PD programs for a total of 30 papers evaluating 33 programs. The search process is detailed in  fig. 2 . The 30 papers are listed in supplementary online appendix A4 .

Search Process and Results for Evaluated Professional Development Programs

Source : Constructed by the authors based on the search described in the text.

Note : The 30 papers documenting the evaluation of the final 33 programs are listed in supplementary online appendix A4 .

Data collection and coding for the sample of 33 evaluated programs comprised two phases. The first of these phases consisted of carefully reviewing the impact evaluation studies and coding the information they provided. The draft version of the instrument for which we collected data included 51 indicators in total, and on average, information on 26 (51 percent) of these indicators was reported in the impact evaluations. Crucially, the amount of program information reported across the impact evaluations varies noticeably by topic ( table 1 ). Sixty-four percent of details concerning the organization of teacher training programs—such as whether the program was designed by a government or by a non-governmental organization (NGO)—can be extracted from the evaluations. In contrast, on average, only 47 percent of information concerning program content and 42 percent of information concerning program delivery is reported.

Data Available on Evaluated Programs from Studies vs. Interviews

Source : Constructed by the authors based on the application of the In-Service Teacher Training Survey Instrument items ( supplementary online appendix A2 ) to the 33 professional development programs identified ( supplementary online appendix A4 ).

Note : Percentage data collected refers to the percentage of indicators for which data were collected across the 33 programs in our evaluated sample. This is calculated by the number of programs for which each indicator has data, summed for every indicator in a given section (or total) and divided by the number of indicators in that section (or total), and finally divided by the 33 programs.

The second phase of data collection sought to fill this gap in reported data by interviewing individuals involved in the actual implementation of each program. To do this, we emailed the authors of each of the impact evaluations in our sample, asking them to connect us with the program implementers. After three attempts to contact the implementers, we received responses from authors for 25 of the 33 programs. We contacted all of the individuals to whom the authors referred us—who in many cases directed us to more relevant counterparts—and were eventually able to hold interviews with program implementers for 18 of the 33 programs. 3 The interviews loosely followed the survey instrument, but included open-ended questions and space for program implementers to provide any additional program information that they perceived as important.

The ITTSI data were gathered retrospectively for this study, which means that in most cases, the evaluation results (and so whether or not the program was effective) were likely to have been known to the interviewee. We propose three reasons that this should not pose a substantive problem for the quality of the data. First, most of the indicators have no normative response. Whether a program is government- or researcher-designed or implemented, whether it has a subject focus or a general pedagogy focus, or whether or not it has a distance learning element have no obvious “right” answers. Second, the survey was administered to program implementers, who usually were not part of the team of researchers who evaluated the program, so they had little stake in confirming research results. Third, the survey had low stakes: interviewees knew that we were independent researchers doing a synthesis review. In some cases, the PD program being discussed no longer existed in the same form. For future PD studies, these data could be collected at the design stage of programs.

For the 18 programs for which we conducted interviews, we were able to collect information for an average of 50 out of the 51 (98 percent) indicators of interest. Consequently, conducting interviews decreased the differences in data availability across categories. The pooled average of indicators for which we had information after conducting interviews (for interviewed and not interviewed programs combined) increased to 79 percent for Organization indicators, 68 percent of Content indicators, and 72 percent of Delivery indicators ( table 1 ).

For our sample of evaluated in-service teacher PD programs, we analyze which characteristics of teacher training programs are associated with the largest improvements in student learning, as measured by test score gains. We conduct both quantitative and qualitative analyses. The analytical strategy for the quantitative analysis essentially consists of comparing means of student learning gains for programs with and without key characteristics, using a bivariate linear regression to derive the magnitude and statistical significance of differences in means. We do not carry out multivariate regression analysis because of the small sample; thus, these results are only suggestive, as multiple characteristics of programs may be correlated. Because we are testing each coefficient separately, we are not able to test the relative value of coefficients, so differences in point estimates are only suggestive.

In preparation for this analysis, we standardize the impact estimates for each of the programs. We convert the program characteristic variables to indicator variables wherever possible to facilitate comparability of coefficients. Although our sample of impact evaluations has a common outcome—impact on student test scores—these are reported on different scales across studies, based on different sample sizes. 4 We standardize these effects and the associated standard errors in order to be able to compare them directly. Supplementary online appendix A5 provides mathematical details of the standardization.

Turning to the independent variables, as originally coded, the 51 indicators for which we collected information capturing various design and implementation characteristics of the PD programs took a number of forms. These consisted of indicator variables (e.g., the intervention provides textbooks alongside training = 0 or 1), categorical variables (e.g., the primary focus of the training was subject content [= 1], pedagogy [= 2], new technology [= 3]), continuous variables (e.g., the proportion of training hours spent practicing with students), and string variables capturing open-ended perceptions (e.g., which program elements do you think were most effective?). To maximize the comparability of output from our regression analysis we convert all categorical and continuous variables into indicator variables. 5

We then conduct our bivariate regressions on this set of complete indicator variables with continuous impact estimates on test scores as the outcome variable for each regression. Because of the limitations associated with running a series of bivariate regressions on a relatively small sample of evaluations, we propose the following robustness check. First, we estimate robust Eicker-Huber-White (EHW) standard errors as our default standard errors (reported in  tables 2 – 4 ) and assess significance according to p -values associated with these. Second, we estimate bootstrapped standard errors and the associated p -values. Third, we run Fisher randomization tests to calculate exact p -values, a common approach in the context of small samples. 6 We report significance under each of these methods separately and report results as robust if they are significant under at least two of the three methods, and if the significant effect is driven by at least two observations—i.e., the results are not explained by a single PD program.

Organization – Bivariate Regressions with Robustness Checks

Source : Constructed by the authors based on data extracted from 33 professional development programs ( supplementary online appendix A4 ) using the In-Service Teacher Training Survey Instrument, and analyzed by regression, as described in the text.

Note : ∗ p  < 0.10, ∗∗ p  < 0.05, ∗∗∗ p  < 0.01 correspond to the significance of p- val ues of robust standard Noteerrors. § corresponds to significance at the 10 percent level or higher for bootstrapped standard errors. † corresponds to significance at the 10 percent level or higher for the Fisher Randomization tests. Numbers specified in parentheses in variable labels are the reported medians for dummy variables in which the variable equals 1 if greater than the median. Total programs refers to the number of programs that report whether or not they have the characteristic. The robust column includes an X if the finding is statistically significant across at least two methods and if the finding is driven by two or more evaluations (i.e., not a single evaluation).

Content – Bivariate Regressions with Robustness Checks

Note : ∗ p  < 0.10, ∗∗ p  < 0.05, ∗∗∗ p  < 0.01 correspond to the significance of p -values of robust standard errors. § corresponds to significance at the 10 percent level or higher for bootstrapped standard errors. † corresponds to significance at the 10 percent level or higher for the Fisher Randomization tests. Total programs refers to the number of programs that report whether or not they have the characteristic. The robust column includes an X if the finding is statistically significant across at least two methods and if the finding is driven by two or more evaluations (i.e., not a single evaluation).

Delivery – Bivariate Regressions with Robustness Checks

Note : ∗ p  < 0.10, ∗∗ p  < 0.05, ∗∗∗ p  < 0.01 correspond to the significance of p -values of robust standard errors. § corresponds to significance at the 10 percent level or higher for bootstrapped standard errors. † corresponds to significance at the 10 percent level or higher for the Fisher Randomization tests. Numbers specified in parentheses in variable labels are the reported medians for dummy variables in which the variable equals 1 if greater than the median. Total programs refers to the number of programs that report whether or not they have the characteristic. The robust column includes an X if the finding is statistically significant across at least two methods and if the finding is driven by two or more evaluations (i.e., not a single evaluation).

We supplement this regression analysis with a qualitative analysis of what works, relying on the self-reported perceptions of program implementers along three dimensions: (a) Which program elements they identified as most responsible for any positive impacts on student learning; (b) which elements, if any, teachers particularly liked; and (c) which elements, if any, teachers particularly disliked.

Applying the ITTSI to At-Scale PD Programs

The sampling process for at-scale programs is detailed in  fig. 3 . To obtain a sample of at-scale, government-funded PD programs across the world, we first identified four to five countries in each region where the World Bank has operations. 7 We worked with regional education managers at the World Bank in each region to select countries in which government counterparts and World Bank country teams had an interest in learning more about in-service teacher PD programs. We made clear that the exercise was appropriate for countries with any level of teacher PD, not specific to countries with recent reforms or innovations. The final set of countries sampled included Burkina Faso, Cambodia, El Salvador, The Gambia, Guinea, India (Bihar state), Jordan, Kazakhstan, the Kyrgyz Republic, Mauritania, Mexico (Guanajato, Oaxaca, and Puebla, and a national PD program for middle school teachers), Moldova, Niger, and the Russian Federation.

Sampling Process for At-Scale Professional Development Programs

Source : Constructed by the authors to reflect the process to identify at-scale professional development programs, as described in the text.

We then obtained permission from the Ministry of Education (MoE) or other relevant government counterparts in each country and worked with them to complete a roster, or listing, of all teacher PD programs conducted between 2012 and 2016. 8 The roster, available in supplementary online appendix A6 , was created along with the ITTSI instrument and collects the following information about each of the teacher PD programs that received government funding: program name; program coordinator's name and contact information; the number of teachers trained; and the types of teachers targeted (e.g., pre-primary, primary, or secondary school teachers). In some countries, such as Mexico and India, where policymaking about teacher PD happens at the state level, we worked with individual states.

After receiving completed roster information about teacher PD programs in a country/state, we used the roster to select a sample of teacher PD programs to interview. In each country/state, we chose the sample by selecting the 10 largest teacher PD programs in terms of teacher coverage, defined as the number of teachers reached by the program during its most recent year of implementation. Of the 10 sampled programs for each country/state, the full ITTSI was administered to the two largest programs targeting primary school teachers and the largest program that targeted secondary school teachers. The brief version of the instrument, the BITTSI, was administered in the remaining seven programs in the country/state. In total, 48 at-scale programs completed the ITTSI and 91 at-scale programs completed the BITTSI across 14 countries.

We applied the ITTSI survey through a combination of phone interviews with and online surveys of PD program coordinators. In a few instances (in The Gambia, El Salvador, and Mexico), depending on the preferences of the program coordinator and their primary language, program coordinators were given the option of completing the ITTSI questionnaire online. For the majority of programs, however, we held phone interviews with program coordinators, in which we asked them the questions included in the ITTSI survey items directly and filled out the instrument ourselves with their responses.

The ITTSI survey applied to the sample of at-scale programs consists of 70 indicators. We were able to collect information for an average of 66 of the 70 (94 percent) indicators of interest for the 48 at-scale teacher PD programs to which the full ITTSI survey was applied, and for 26.5 of the 27 (97 percent) indicators—derived from 13 questions—for the 91 programs to which the BITTSI was applied.

For the sample of at-scale PD programs, we compare the average of observed characteristics of at-scale teacher PD programs with the average for evaluated PD programs that resulted in the largest improvements in student learning (“top performers”), as measured by student test score gains. To determine the characteristics of “top performers,” we ranked all evaluated programs, using their standardized impact on student test scores. We then selected the top half of programs (16 programs, all of which displayed positive impacts), and calculated the average value of program indicators for those “top performers.” We compare them to the means of at-scale PD programs in order to better understand the gap between at-scale PD practices and the best practices of top-performing PD programs.

This section characterizes the specific characteristics of teacher PD programs that successfully improve student learning in low- and middle-income countries and how common these characteristics are across at-scale, government-funded programs. First, we present the results of our quantitative and qualitative analyses examining which PD characteristics are associated with large gains in student learning for the sample of evaluated programs. Second, we present descriptive statistics from the sample of at-scale PD programs and from the top-performing PD programs in the evaluated sample to shed light on how they differ in terms of those PD characteristics found to be associated with positive impacts on student learning.

Which PD Characteristics are Most Associated with Student Learning Among Evaluated Programs?

We discuss, for each of our categories—Organization, Content, and Delivery—those characteristics we observe to be most associated with student learning gains.  Tables 2 – 4 present the results of our bivariate regressions for each of these categories in turn. In each case, we report the results with the three different methods of calculating significance as well as an indicator of robustness.

Among Organization ( table 2 ), two characteristics are robustly associated with significant gains in student learning. These include linking career opportunities (improved status, promotion, or salary) to PD programs and targeting training programs based on teachers’ years of experience. First, in teacher PD programs where participation has no implications for promotion, salary, or status increases, student learning is 0.12 standard deviations lower (significant at 95 percent). In other words, programs that do link participation to career incentives have higher effectiveness. 9 Second, targeting participant teachers by their years of experience is associated with 0.10 standard deviations higher student learning (significant at 90 percent). This is driven by two programs: the Balsakhi program in rural India, which trains women from the local community who have completed secondary school to provide remedial education to students falling behind ( Banerjee et al. 2007 ); and the Science teacher training program in Argentina, which trains teachers in different structured curricula and coaching techniques and finds that coaching is only effective for less experienced teachers ( Albornoz et al. 2018 ). Indeed, these are the only two programs out of the 33 that explicitly targeted teachers based on their experience, both of which resulted in student learning gains. In addition, the provision of complementary materials such as storybooks and other reading materials (e.g., flashcards or word banks) have large coefficients associated with improving student learning (0.11 and 0.13 standard deviations), although these are not statistically significant.

Among the Content variables ( table 3 ), programs with a specific subject focus result in higher learning gains than more general programs. Specifically, programs with no subject focus show 0.24 standard deviations lower impact on student learning (significant at 99 percent). A deeper look reveals that within focus areas, programs that are not focused on a given academic subject—such as those focused on counseling—are associated with 0.2 lower standard deviations in student learning (significant at 99 percent). Lastly, when a teacher PD program involves teaching practice through lesson enactment, it is associated with a 0.10 standard deviation increase in student learning (significant at 90 percent).

Turning to Delivery characteristics ( table 4 ), three characteristics of teacher PD programs are robust. First, teacher PD programs that provide consecutive days of face-to-face teacher training are associated with a 0.14 standard deviation increase in student learning (significant at 99 percent). Second, holding face-to-face training at a central location—such as a hotel or government administrative building (as opposed to a university or training center, which was the omitted category)—is associated with a 0.13 lower standard deviation in student learning (significant at 90 percent). Third, teacher PD training programs that are conducted remotely using distance learning are associated with a 0.10 standard deviation decrease in student learning (significant at 90 percent). In alignment with recent literature highlighting the overly theoretical nature of many training programs as an explanation for their limited effects on student learning—as well as the above finding that training programs that involve teaching practice are associated with 0.16 larger gains in student learning—the proportion of training time spent practicing with other teachers is highly correlated with learning impacts (although not consistently statistically significant). Also, the inclusion of follow-up visits to review material taught in the initial training—as opposed to visits for monitoring purposes alone or no follow-up visits—is associated with a 0.14 standard deviation higher program impact on student learning (not significant, but one of the largest coefficients). These findings support the literature that subject-focused teacher PD programs with consecutive days of face-to-face training that include time for teachers to practice with one another, are associated with improved student learning outcomes.

We supplement the quantitative results with an analysis of self-reported perceptions by the implementers of the evaluated programs. These concern the characteristics of their programs which they believe are most responsible for any positive effects on student learning, as well as those elements which were popular and unpopular among the beneficiary teachers. We elicited these perceptions using open-ended questions and then tallied the number of program implementers that mentioned a given program element in their response, albeit not necessarily using the exact same language as other respondents. These responses come from 18 interviewees, so they should be taken as suggestive. That said, the results broadly align with the quantitative results: Five of 18 interviewees—tied for the most common response—mentioned that mentoring follow-up visits were a crucial component in making their training work. Similarly, five of the 18 interviewees discuss the importance of having complementary materials, such as structured lessons or scripted materials that provide useful references in the classroom and help to guide teachers during the training sessions. The next most commonly reported elements were engaging teachers for their opinions and ideas—either through discussion or text messages—and designing the program in response to local context, building on what teachers already do and linking to everyday experiences: both were mentioned by four of 18 interviewees.

We also asked the program implementers about the program characteristics that they believed teachers liked and disliked the most about their training programs and, interestingly, we only found two common responses for what teachers particularly liked and one common response for what they disliked. 10 Seven of the 18 interviewees reported that the part of their program that teachers most enjoyed was that it was fun and engaging (or some variation of that). In other words, teachers appreciated that certain programs were interactive and involved participation and discussion rather than passive learning. In addition to having “fun” teacher PD programs, five of the 18 interviewees suggested that teachers especially liked the program materials provided to them. Similarly, in terms of unpopular program elements, four of the 18 program implementers we interviewed reported that teachers disliked the amount of time taken by participating in the training programs, which they perceived as excessive.

What Do We Learn from At-Scale PD Programs?

Government-funded, at-scale teacher PD programs have a number of characteristics in common ( supplementary online appendix tables A7.1–A7.3 ). The vast majority are designed by government (80 percent) and implemented by government (90 percent). Almost all provide materials to accompany the PD (96 percent), and most include at least some lesson enactment (73 percent) and development of materials (73 percent). Most have a subject focus (92 percent) and include an initial period of face-to-face training for several days (85 percent). Most do not formally target teachers by subject (only 19 percent do), grade (31 percent), or years of experience (13 percent), and few have negative consequences if teachers are poorly evaluated (17 percent). These at-scale programs differ sharply from programs that are evaluated in general, as well as from top-performing evaluated programs specifically. We provide a full list of average characteristics of at-scale programs and all evaluated programs (not just top-performers) in supplementary online appendix tables A7.1–A7.3 .

Our principal focus in this section is how at-scale programs compare to evaluated programs that deliver relatively high gains in student learning. We assess the top half of programs (N = 16) from the sample of evaluated programs by selecting those characteristics that produced the largest standard deviation increases in student assessment scores. In  table 5 , we compare the means of at-scale programs and top-performing, evaluated programs. We focus specifically on the characteristics shown to have a statistically significant relationship with student learning outcomes and those with large coefficients, identified for interest (as identified in  tables 2 – 4 ).

Comparison of Means of At-Scale Programs and Top-Performing, Evaluated Programs

Source : Constructed by authors, comparing summary statistics for the top performing professional development (PD) programs among rigorously evaluated PD programs to at-scale PD programs.

Note : For the full list of statistics, see supplementary online appendix Tables A7.1–A7.3 .

Regarding Organization ( table 5 ), two key characteristics—whether or not the training is linked to career opportunities and whether or not the program targets teachers based on their years of experience—are robustly associated with improved student learning gains. There are notable and substantive differences between top-performing PD programs and the sample of at-scale PD programs when it comes to providing incentives; 88 percent of top-performing PD programs link training to status or to new career opportunities such as promotion or salary, as compared to only 55 percent of at-scale programs. Our results suggest that without incentives, training may not have a meaningful impact. Furthermore, top-performing programs and at-scale PD programs are similar in the degree to which they target teachers based on their years of experience. For instance, 13.3 percent of top-performers and 12.5 percent of at-scale programs target teachers based on their experience. Other notable organizational characteristics include the provision of complementary materials such as storybooks and reading materials. Top-performing PD programs and at-scale PD programs are similar in the amount of materials they provide, but our results suggest that the kinds of complementary materials may differ somewhat. For instance, only 12.5 percent and 21 percent of at-scale programs provide storybooks and reading materials, respectively—materials correlated with student learning gains—as compared to 36 percent and 43 percent of evaluated programs.

Turning next to Content ( table 5 ), top-performing PD programs and at-scale PD programs perform similarly. In both instances, the majority of programs include subject content and subject-specific pedagogy as either a primary or secondary focus. Few programs—none of the top performers—and only eight percent of at-scale programs lack a subject focus. Moreover, no top-performing programs and few at-scale programs (fewer than six percent) focus on general training in areas such as counseling or providing training on how to use a specific tool—types of training that are statistically linked to lower gains in student learning.

Finally, Delivery characteristics ( table 5 ) include whether or not there are consecutive days of face-to-face training, training location, the amount of time teachers spend practicing with one another, and follow-up visits. Specifically, 100 percent of top-performing programs include consecutive days of face-to-face training as compared to 85 percent of evaluated programs. Our research further suggests that the location of PD training programs may influence program effectiveness, and training held at central locations such as hotels or conference rooms (as opposed to universities or training centers) may be less effective. Currently 73 percent of at-scale, government-funded programs are held at central locations as compared to only 38 percent of evaluated programs.

Follow-up visits with teachers and the amount of time teachers spend practicing with other teachers during the training program are shown to be positively correlated with large coefficients (albeit not statistically significant) on student learning. In both instances, top-performing PD programs include more follow-up visits (five versus two median visits among programs with visits) and spend more time allowing teachers to practice with other teachers (40 percent versus 16 percent of training time) than do at-scale programs. 11 Results of our analysis suggest that training may be more effective if there are follow-up visits. This is an imperative finding when comparing top-performing PD programs, in which 85 percent include follow-up visits, with government-funded, at-scale PD programs, in which only half of programs include follow-up visits. Also, in top-performing PD programs, teachers spend more time practicing what they have learned with other teachers (40 percent of overall training time) relative to at-scale programs (only 16 percent). An existing body of research suggests that when teachers have opportunities to practice the new skills they acquire in PD programs, they are more likely to adopt these new skills in their classrooms ( Wiley and Yoon 1995 ; Wenglinsky 2000 ; Angrist and Lavy 2001 ; Borko 2004 ).

Governments spend enormous amounts of time and money on in-service professional development. Many countries have multiple in-service PD programs running simultaneously, as evidenced by our sample of at-scale PD programs. Many go unevaluated and may be ineffective. This paper makes three major contributions: first, it reveals broad weaknesses in reporting on teacher PD interventions. There are almost as many program types as there are programs, with variations in subject and pedagogical focus, hours spent, capacity of the trainers, and a host of other variables. Yet reporting on these often seeks to reduce them to a small handful of variables, and each scholar decides independently which variables are most relevant to report. We propose a standard set of indicators—the ITTSI—that would encourage consistency and thoroughness in reporting. Academic journals may continue to pressure authors to report limited information about the interventions, wishing instead to reserve space for statistical analysis. However, authors could easily include the full set of indicators in an appendix attached to the paper or online.

Second, this paper demonstrates that some characteristics of teacher PD programs—notably, linking participation to incentives such as promotion or salary implications, having a specific subject focus, incorporating lesson enactment in the training, and including initial face-to-face training—are positively associated with student test score gains. Furthermore, qualitative evidence suggests that follow-up visits to reinforce skills learned in training are important to effective training. Further documentation of detailed program characteristics, coupled with rigorous evaluation, will continue to inform effective evaluations.

The impacts of these characteristics are not small: having a specific subject focus and incorporating lesson enactment are associated with 0.24 and 0.10 more standard deviations in learning, respectively, for example. Comparing these effect sizes to those from a sample of 747 education-related randomized controlled trials in the United States puts them both above the 50th percentile in terms of effectiveness ( Kraft 2020 ). Comparing to a set of 130 randomized controlled trials in low- and middle-income countries likewise put them at or above the 50 th percentile of 0.10 standard deviations ( Evans and Yuan 2020 ). In high-income countries, Kennedy (2019) proposes that the impact of teacher PD programs be benchmarked against a much less costly “community of practice” model in which teachers help each other, like Papay et al. (2020) . While we are not aware of a rigorously evaluated, costed model of that class of program in a low- or middle-income country, an alternative would be to compare teacher PD results to a pure monitoring model, such as an increase in inspections. Along these lines, Muralidharan et al. (2017) show—using data from India—that increased frequency of monitoring would be a much more cost-effective way to reduce effective class sizes (through reduced teacher absenteeism) than hiring more teachers. These are useful avenues to pursue for future research as countries consider the cost-effectiveness of alternative investments in teachers.

Third, by comparing the means of at-scale PD programs with top-performing evaluated programs, our findings highlight gaps between what evidence suggests are effective characteristics of teacher PD programs and the contextual realities of most teacher PD programs in their design, content, and delivery. In particular, our findings taken together suggest that at-scale programs often lack key characteristics of top-performing training programs. At-scale programs are much less likely to be linked to career incentives, to provide storybooks or other reading materials, to have a subject content focus, to include time for practicing with other teachers, or to include follow-up visits.

The approach taken by this paper centers on using the ITTSI to collect and compare data on rigorously evaluated and at-scale, government-funded teacher PD programs. This approach has limitations. First, the evidence of what works within rigorously evaluated programs is limited by those programs that have been evaluated. There may be innovative PD programs that are not among the “top performers” simply because they have yet to be evaluated. While this evidence base can push policymakers away from approaches that do not work, it should not deter policymakers from innovating and evaluating those innovations.

A second, related limitation concerns the relatively small sample of evaluated teacher PD programs in low- and middle-income countries, on which our findings about effective PD characteristics are based. Some of the larger coefficients in the regressions are driven by a small number of teacher training programs. These instances have been noted in the text. As more evaluations of PD programs are conducted, the ITTSI can be applied to these and our analyses re-run to shed further light on the specific characteristics associated with PD programs that improve student learning. The ITTSI data were already updated once in this way in 2018, increasing the number of evaluated programs in our sample from 26 to 33.

Third, a conceptual concern with evaluating teacher professional programs is the risk that impacts may be explained by observer effects (also referred to as Hawthorne effects). These effects have been documented in education ( Muralidharan et al. 2017 ) and health in low- and middle-income countries ( Leonard 2008 ; Leonard and Masatu 2010 ). The impact of any education intervention may partly be due to observer effects, since the introduction of an intervention suggests that someone is paying attention to the teacher's efforts. Both randomized controlled trials and more traditional monitoring and evaluation may enhance these effects, as teachers may further respond favorably to the observation associated with measurement. Randomized controlled trials and quasi-experimental studies with a credible comparison group overcome part of this concern, as the observer effect associated with measurement will exist in both the treatment and comparison groups, and measured program impacts should be net of those effects.

That leaves the impact of the intervention itself. In this review, all of the studies we include evaluate interventions and, as such, all may be subject to an observer effect. Our analysis implicitly assumes the magnitude of this observer effect to be constant across different types of PD. By comparing PD characteristics across programs, we observe whether those characteristics are associated with a larger total effect on learning. Part of that total effect may stem from increased teacher skills, and part may be explained by certain PD characteristics inducing greater observer effects (since any observer effects that are uncorrelated with PD characteristics would be absorbed in our regression constant terms). In the short run, the impact for students is observationally equivalent. Even with longer run studies (of which there are very few in education and development), observer effects may fade, but teacher skills may also depreciate ( Cilliers et al. 2020 ). As a result, we consider the total association of PD characteristics with student learning, including through increased teacher human capital and observer effects.

Fourth, there are challenges in comparing evaluated PD programs with at-scale PD programs. As the data demonstrate, at-scale PD programs tend to be larger programs designed by governments, often at the national level, and aimed at providing broad training to teachers. In light of these differences, we highlight the fact that top-performing programs—regardless of their core objectives—share certain common sets of characteristics that most at-scale programs do not share. Awareness of these characteristics may be useful in the conceptualization and implementation of future teacher PD programs in low- and middle-income countries, including large-scale programs funded by governments.

One key reason that at-scale programs may differ from successful, evaluated programs is that the latter group of evaluations may not be designed in a way that is conducive to scaling. Evaluated programs tend to be much smaller than at-scale programs: in our data, evaluated programs reached an average of 96 schools versus at-scale programs that reached more than 6,000 schools on average ( supplementary online appendix table A7.1 ). These smaller programs often have higher per-pupil costs ( Evans and Popova 2016b ), so scaling them nationwide requires cutting elements. Smaller programs are easier to staff and easier to monitor. Evaluated programs were three times as likely to be designed by researchers and less than one-third as likely to be implemented by government ( supplementary online appendix table A7.1 ). One solution, obviously, is more large-scale evaluations, like Loyalka et al. (2019) . However, even smaller evaluations can do more to mimic scalable policies. Gove et al. (2017) , reflecting on programs evaluated both at pilot and at scale in Kenya and Liberia, suggest the value of testing as many elements as possible in the pilot, using government systems in the pilot as much as possible, and to make sure that pilot costs are within what a government budget can handle. Duflo et al. (2020) combine these two approaches in a recent nationwide, five-arm randomized controlled trial in Ghana, to test the scalability of four different models to reach remedial learners, which had previously been tested in small pilot randomized controlled trials elsewhere. When implemented within existing government systems, they find all four interventions to be effective, pointing to the program's inception within the government as key, as opposed to an initial non-government organization initiative subsequently and imperfectly implemented by the government.

Improving in-service teacher professional development may be a clear win for governments. They are already spending resources on these programs, and there is broad support for these programs among teachers and teachers’ unions. Interventions such as the above provide learning opportunities for country governments and stakeholders seeking to design effective teacher PD programs. While no single characteristic of top-performing PD programs may transform an ineffective PD program into an effective one, this paper highlights trends in top-performing programs, such as including incentives, a specific subject focus, and lesson enactment. These are characteristics that, if included and implemented successfully, have the potential to improve the quality of teacher PD programs, and ultimately, the quality of instruction and student learning.

The authors are grateful for comments from Denise Bello, Luis Benveniste, Barbara Bruns, Martin Carnoy, Joost de Laat, Margaret Dubeck, Deon Filmer, Susanna Loeb, Prashant Loyalka, Ezequiel Molina, Andrew Ragatz, and Halsey Rogers. They are also grateful to Fei Yuan for excellent research assistance, to Veronica Michel Gutierrez, Olga A. Rines, Lea Jeanne Marie Lungmann, Fata No, and Elissar Tatum Harati for their support with data collection, and to numerous teacher training implementers for providing information on programs. This paper subsumes an earlier paper, “Training Teachers on the Job: What Works and How to Measure It” (World Bank Policy Research Working Paper Number 7834).

This work was supported by the Bill & Melinda Gates Foundation, the World Bank's Systems Approach for Better Education (SABER) Trust Fund, which was supported by the United Kingdom's Department for International Development (DFID) and Australia's Department of Foreign Affairs and Trade (DFAT), and the Strategic Impact Evaluation Fund at the World Bank.

Both samples focus on teacher training programs at the primary and secondary school level. Pre-primary schools are excluded.

The databases we searched were the Education Resources Information Center (ERIC); Academic Search Complete; Business Source Complete; Econlit with Full Text; Education Full Text (H. W. Wilson); Education Index Retrospective: 1929–1983; Education Source; Educational Administration Abstracts; Social Science Full Text (H. W. Wilson); Teacher Reference Center; and EconLit. We looked for articles containing the terms (“teacher training” OR “teacher education” OR “professional development”) AND (``learning'' OR ``scores'' OR ``attainment'') AND (“impact evaluation” OR ``effects'') AND (“developing country 1” OR “developing country 2” OR “developing country N”), where “developing country” was replaced by country names.

In six cases, program implementers failed to schedule an interview after three attempts at contact, and in the case of one older program, the implementer had passed away. Interviews were held over the phone or in-person, and lasted between 45 and 90 minutes for each program.

A limitation is that some of the impact estimates from school-randomized control trials in our evaluated sample are over-estimates because the authors fail to account for the clustering of children within teachers or schools ( Hedges 2009 ).

For categorical variables, this is straightforward. For example, we convert the original categorical variable for the location of the initial teacher PD—which includes response options of schools, a central location, a training center, or online—into four dummy variables. In order to convert the continuous variables to a comparable scale, we create a dummy for each continuous variable which, for a given program, takes a value of 1 if the continuous variable is greater than the median value of this variable across all programs, and a value of 0 if it is less than or equal to the value of this variable across all programs. We apply this method to the conversion of all continuous variables except three—proportion of teachers that dropped out of the program, number of follow-up visits, and weeks of distance learning—which we convert directly to dummy variables that take a value of 1 if the original variable was greater than 0, and a value of 0 otherwise.

We estimate bootstrapped standard errors by resampling our data with replacement 1,000 times. We run Fisher randomization tests by treating each indicator PD characteristic as a treatment and calculating a randomization distribution of mean differences (the test statistic) across treatment assignments. Specifically, for 1,000 permutations, we randomly reassign values of 0 or 1 to the independent variables in our regressions, while maintaining the overall proportion of 0s and 1s observed in the empirical sample for a given variable. We then calculate Fisher exact p -values by finding the proportion of the randomization distribution that is larger than our observed test statistic ( Fisher 1925 , 1935 ; Imbens and Rubin 2015 ).

These regions include: Africa, Eastern and Central Europe, Latin American and the Caribbean, the Middle East and North Africa, and East and South Asia.

This includes programs ongoing in 2016 and programs that were implemented anytime in the range of 2012 to 2016. Hence, the programs could have been designed prior to 2012. We still include them if they were implemented any time between 2012 and 2016. We were not successful in obtaining roster information in all countries. For instance, in Morocco and the Arab Republic of Egypt, the Ministries of Education were in the process of making changes to the structure and delivery of teacher training programs and indicated that it was not a good time for data collection. In Tanzania there was a change in leadership among government counterparts during efforts to complete the roster and data collection process, and we were not able to properly sample and apply the ITTSI in all teacher-training programs in the country. In India, we had initially identified two states, Bihar and Karnataka, to work with at the subnational level, but ultimately only collected data in one state, Bihar, since the principal government counterpart in Karnataka was not available to complete the roster.

In some cases, we test a negative (e.g., no implications for status in table 2 or no subject focus in table 3 ) because we are testing an exhaustive series of indicators derived from the same question (e.g., subject focus is math, subject focus is literacy, or no subject focus).

Because it is difficult to imagine an effective teacher professional development program that teachers actively dislike (they have to learn for it to work, after all), their preferences are relevant.

When we include programs with no follow-up visits, the median number of follow-up visits to teachers in top programs becomes 3.5 as compared to 0 for at-scale programs.

Albornoz   F. , Anauati   M. V. , Furman   M. , Luzuriaga   M. , Podestá   M. E. , Taylor   I. . 2018. “ Training to Teach Science: Experimental Evidence from Argentina .” Policy Research Working Paper 8594 , World Bank, Washington, DC .

Angrist   J. D. , Lavy   V. . 2001 . “ Does Teacher Training Affect Pupil Learning? Evidence from Matched Comparisons in Jerusalem Public Schools .” Journal of Labor Economics   19 ( 2 ): 343 – 69 .

Google Scholar

Araujo   M. C. , Carneiro   P. , Cruz-Aguayo   Y. , Schady   N. . 2016 . “ Teacher Quality and Learning Outcomes in Kindergarten .” Quarterly Journal of Economics   131 ( 3 ): 1415 – 53 .

Baker   S. , Smith   S. . 1999 . “ Starting off on the Right Foot: The Influence of Four Principles of Professional Development in Improving Literacy Instruction in Two Kindergarten Programs .” Learning Disabilities Research & Practice   14 ( 4 ): 239 – 53 .

Banerjee   A. , Cole   S. , Duflo   E. , Linden   L. L. . 2007 . “ Remedying Education: Evidence from Two Randomized Experiments in India .” Quarterly Journal of Economics   122 ( 3 ): 1235 – 64 .

Berlinski   S. , Busso   M. . 2017 . “ Challenges in Educational Reform: An Experiment on Active Learning in Mathematics .” Economics Letters   156 : 172 – 5 .

Bold   T. , Filmer   D. , Martin   G. , Molina   E. , Rockmore   C. , Stacy   B. , Svensson   J. , Wane   W. . 2017 . “ What Do Teachers Know and Do? Does It Matter? Evidence from Primary Schools in Africa .” Policy Research Working Paper 7956 , World Bank, Washington, DC .

Borenstein   M. , Hedges   L. V. , Higgins   J. P. T. , Rothstein   H. R. . 2009 . Introduction to Meta-Analysis .   Chichester, United Kingdom : John Wiley & Sons .

Google Preview

Borko   H.   2004 . “ Professional Development and Teacher Learning: Mapping the Terrain .” Educational Researcher   33 ( 8 ): 3 – 15 .

Bourgeois   E. , Nizet   J. . 1997 . Aprendizaje y formación de personas adultas .   Paris, France : Presses Universite de France .

Cardemil   C.   2001 . “ Procesos y condiciones en el aprendizaje de adultos .” Jornada Nacional de Supervisores. Supervisión para aprendizajes de calidad y oportunidades para todos. Educación Rural .   Santiago : Ministerio de Educación . https://repositorio.uahurtado.cl/handle/11242/8517 .

Chetty   R. , Friedman   J. N. , Rockoff   J. E. . 2014 . “ Measuring the Impacts of Teachers II: Teacher Value-Added and Student Outcomes in Adulthood .” American Economic Review   104 ( 9 ): 2633 – 79 .

Cilliers   J. , Fleisch   B. , Kotze   J. , Mohohlwane   M. , Taylor.   S.   2020 . “ The Challenge of Sustaining Effective Teaching: Spillovers, Fade-out, and the Cost-effectiveness of Teacher Development Programs .” Unpublished Working Paper . https://www.dropbox.com/s/6xmv7283oxoysj2/The%20Challenge%20of%20Sustaining%20Effective%20Teaching%20with%20appendix.pdf?dl=0 .

Darling-Hammond   L. , Wei   R. C. , Andree   A. , Richardson   N. , Orphanos   S. . 2009 . Professional Learning in the Learning Profession .   Washington, DC : National Staff Development Council . https://edpolicy.stanford.edu/sites/default/files/publications/professional-learning-learning-profession-status-report-teacher-development-us-and-abroad.pdf .

Desimone   L. M.   2009 . “ Improving Impact Studies of Teachers’ Professional Development: Toward Better Conceptualizations and Measures .” Educational Researcher   38 ( 3 ): 181 – 99 .

Duflo   A. , Kiessel   J. , Lucas   A. . 2020 . “ External Validity: Four Models of Improving Student Achievement .” Working Paper No. w27298 , National Bureau of Economic Research , Cambridge, MA .

Evans   D. K. , Popova   A. . 2016a . “ What Really Works to Improve Learning in Developing Countries? An Analysis of Divergent Findings in Systematic Reviews .” World Bank Research Observer   31 ( 3 ): 242 – 70 .

Evans   D. K. , Popova   A. . 2016b . “ Cost-Effectiveness Analysis in Development: Accounting for Local Costs and Noisy Impacts .” World Development   77 : 262 – 76 .

Evans   D. K. , Yuan   F. . 2020 . “ How Big are Effect Sizes in International Education Studies? ” Working Paper 545 , Center for Global Development , Washington, DC .

Fisher   R. A.   1925 . Statistical Methods for Research Workers , first edition. Edinburgh : Oliver and Boyd Ltd .

Fisher   R. A.   1935 . The Design of Experiments ,  sixth edition. Edinburgh : Oliver and Boyd, Ltd , 1951 .

Fryer, Jr   R. G.   2017 . “ The Production of Human Capital in Developed Countries: Evidence from 196 Randomized Field Experiments .” Handbook of Economic Field Experiments   2 : 95 – 322 .

Gersten   R. , Taylor   M. J. , Keys   T. D. , Rolfhus   E. , Newman-Gonchar   R. . 2014 . Summary of Research on the Effectiveness of Math Professional Development Approaches .   Washington, DC : Institute of Education Sciences, National Center for Education Evaluation and Regional Assistance, U.S. Department of Education; and Regional Educational Laboratory Southeast at Florida State University . http://files.eric.ed.gov/fulltext/ED544681.pdf .

Gove   A. , Poole   M. K. , Piper   B. . 2017 . “ Designing for Scale: Reflections on Rolling Out Reading Improvement in Kenya and Liberia .” New Directions for Child and Adolescent Development   2017 ( 155 ): 77 – 95 .

Hedges   L.V . 2009 . “ Effect Sizes in Nested Designs .” In The Handbook of Research Synthesis and Meta-analysis , edited by Cooper   H. , Hedges   L. V. , Valentine,   J. C.   337 – 56 . New York, NY : Russell Sage Foundation .

Huberman   M.   1989 . “ The Professional Life Cycle of Teachers .” Teachers College Record   91 ( 1 ): 31 – 57 .

Imbens   G. W. , Rubin   D. B. . 2015 . Causal Inference in Statistics, Social, and Biomedical Sciences .   Cambridge, UK: Cambridge University Press .

Kennedy   M. M . 2019 . “ How We Learn About Teacher Learning .” Review of Research in Education   43 ( 1 ): 138 – 62 .

Kerwin   J. T. , Thornton   R. L. . 2021 . “ Making the Grade: The Sensitivity of Education Program Effectiveness to Input Choices and Outcome Measures .” Review of Economics and Statistics   103 ( 2 ): 251 – 64 .

Knowles   M. S. , Holton   E. F. , Swanson   R. A. . 2005 . The Adult Learner , sixth edition.   Burlington, MA : Elsevier .

Kraft   M. A.   2020 . “ Interpreting Effect Sizes of Education Interventions .” Educational Researcher   49 ( 4 ): 241 – 53 .

Kraft   M. A. , Papay   J. P. . 2014 . “ Can Professional Environments in Schools Promote Teacher Development? Explaining Heterogeneity in Returns to Teaching Experience .” Educational Evaluation and Policy Analysis   36 ( 4 ): 476 – 500 .

La Paro   K. M. , Pianta   R. C. . 2003 . CLASS: Classroom Assessment Scoring System . Charlottesville, VA : University of Virginia .

Leonard   K. L.   2008 . “ Is Patient Satisfaction Sensitive to Changes in the Quality of Care? An Exploitation of the Hawthorne Effect .” Journal of Health Economics   27 ( 2 ): 444 – 59 .

Leonard   K. L. , Masatu   M. C. . 2010 . “ Using the Hawthorne Effect to Examine the Gap Between a Doctor's Best Possible Practice and Actual Performance .” Journal of Development Economics   93 ( 2 ): 226 – 34 .

Loyalka   P. , Popova   A. , Li   G. , Shi   Z. . 2019 . “ Does Teacher Training Actually Work? Evidence from a Large-Scale Randomized Evaluation of a National Teacher Training Program .” American Economic Journal: Applied Economics   11 ( 3 ): 128 – 54 .

McEwan   P.   2015 . “ Improving Learning in Primary Schools of Developing Countries: A Meta-analysis of Randomized Experiments .” Review of Educational Research   85 ( 3 ): 353 – 94 .

Molina   E. , Fatima   S. F. , Ho   A. , Hurtado   C. M. , Wilichowski   T. , Pushparatnam   A. . 2018 . “ Measuring Teaching Practices at Scale: Results from the Development and Validation of the Teach Classroom Observation Tool .” Policy Research Working Paper 8653 , World Bank , Washington, DC .

Muralidharan   K. , Das   J. , Holla   A. , Mohpal   A. . 2017 . “ The Fiscal Cost of Weak Governance: Evidence from Teacher Absence in India .” Journal of Public Economics   145 : 116 – 35 .

Papay   J. P. , Taylor   E. S. , Tyler   J. H. , Laski   M. E. . 2020 . “ Learning Job Skills from Colleagues at Work: Evidence from a Field Experiment Using Teacher Performance Data .” American Economic Journal: Economic Policy   12 ( 1 ): 359 – 88 .

Piper   B. , Korda   M. . 2011 . EGRA Plus: Liberia (Program evaluation report) . Durham, NC : RTI International .

Rogers   H. , Demas   A. . 2013 . The What, Why, and How of the Systems Approach for Better Education Results (SABER) . Washington, DC: World Bank . http://wbgfiles.worldbank.org/documents/hdn/ed/saber/supporting_doc/Background/SABER_Overview_Paper.pdf .

Shulman   L. S.   1986 . “ Those Who Understand: Knowledge Growth in Teaching .” Educational Researcher   15 ( 2 ): 4 – 14 .

TNTP . 2015 . The Mirage: Confronting the Hard Truth about Our Quest for Teacher Development . The New Teacher Project. http://files.eric.ed.gov/fulltext/ED558206.pdf .

Villegas-Reimers   E.   2003 . Teacher Professional Development: An International Review of the Literature . Paris : UNESCO International Institute for Educational Planning . http://www.iiep.unesco.org/en/publication/teacher-professional-development-international-review-literature .

Walter   C. , Briggs   J. . 2012 . What Professional Development Makes the Most Difference to Teachers .   Oxford : University of Oxford Department of Education . https://www.oupjapan.co.jp/sites/default/files/contents/events/od2018/media/od18_Walter_reference.pdf .

Wenglinsky   H.   2000 . “ How Teaching Matters: Bringing the Classroom Back into Discussions of Teacher Quality .” Policy Information Center Report , Educational Testing Service (ETS) .

Wiley   D. , Yoon   B. . 1995 . “ Teacher Reports of Opportunity to Learn: Analyses of the 1993 California Learning Assessment System .” Educational Evaluation and Policy Analysis   17 ( 3 ): 355 – 70 .

Wood   F. H. , McQuarrie   F. Jr.   1999 . “On the Job Learning. New Approaches will Shape Professional Learning in the 21st Century .” Journal of Staff Development   20 : 10 – 13 .

Yoon   K. S. , Duncan   T. , Lee   S. W. Y. , Scarloss   B. , Shapley   K. . 2007 . Reviewing the Evidence on how Teacher Professional Development Affects Student Achievement (Issues & Answers Report No. 033) . Washington, DC : Institute of Education Sciences, National Center for Education Evaluation and Regional Assistance, U.S. Department of Education; and Regional Educational Laboratory Southeast at Florida State University . http://files.eric.ed.gov/fulltext/ED498548.pdf .

Supplementary data

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1564-6971
• Print ISSN 0257-3032
• Copyright © 2024 World Bank
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Effective Teacher Professional Development

• Press Release

Educators and policymakers are increasingly looking to teacher professional learning as an important strategy for supporting the complex skills students need to be prepared for further education and work in the 21st century. For students to develop mastery of challenging content, problem-solving, effective communication and collaboration, and self-direction, teachers must employ more sophisticated forms of teaching. Effective professional development (PD) is key to teachers learning and refining the pedagogies required to teach these skills.

But what constitutes effective professional development? That’s the question we set out to answer in this report, which reviews 35 methodologically rigorous studies that have demonstrated a positive link between teacher professional development, teaching practices, and student outcomes. We identify key features of effective efforts and offer rich descriptions of these models to inform education leaders and policymakers seeking to leverage professional development to improve student learning.

Defining and Studying Effective Professional Development

We define effective professional development as structured professional learning that results in changes in teacher practices and improvements in student learning outcomes. To define features of effective PD, we reviewed studies meeting our methodological criteria that emerged from our extensive search of the literature over the last three decades.

Using this methodology, we found that effective professional development incorporates most, if not all, of the following elements:

• Is content focused: PD that focuses on teaching strategies associated with specific curriculum content supports teacher learning within teachers’ classroom contexts. This element includes an intentional focus on discipline-specific curriculum development and pedagogies in areas such as mathematics, science, or literacy.
• Incorporates active learning:  Active learning engages teachers directly in designing and trying out teaching strategies, providing them an opportunity to engage in the same style of learning they are designing for their students. Such PD uses authentic artifacts, interactive activities, and other strategies to provide deeply embedded, highly contextualized professional learning. This approach moves away from traditional learning models and environments that are lecture based and have no direct connection to teachers’ classrooms and students.
• Supports collaboration: High-quality PD creates space for teachers to share ideas and collaborate in their learning, often in job-embedded contexts. By working collaboratively, teachers can create communities that positively change the culture and instruction of their entire grade level, department, school and/or district.
• Uses models of effective practice:  Curricular models and modeling of instruction provide teachers with a clear vision of what best practices look like. Teachers may view models that include lesson plans, unit plans, sample student work, observations of peer teachers, and video or written cases of teaching.
• Provides coaching and expert support: Coaching and expert support involve the sharing of expertise about content and evidence-based practices, focused directly on teachers’ individual needs.
• Offers feedback and reflection: High-quality professional learning frequently provides built-in time for teachers to think about, receive input on, and make changes to their practice by facilitating reflection and soliciting feedback. Feedback and reflection both help teachers to thoughtfully move toward the expert visions of practice.
• Is of sustained duration: Effective PD provides teachers with adequate time to learn, practice, implement, and reflect upon new strategies that facilitate changes in their practice.

The report also examines professional learning communities (PLCs) as an example of a PD model that incorporates several of these effective elements and supports student learning gains. This collaborative and job-embedded PD can be a source of efficacy and confidence for teachers, and can result in widespread improvement within and beyond the school level.

Creating Conditions for Effective Professional Development: Opportunities and Challenges

Research has established that the educational system within which PD occurs has implications for its effectiveness. Specifically, conditions for teaching and learning both within schools and at the system level can inhibit the effectiveness of PD. For example, inadequate resourcing for PD—including needed curriculum materials—frequently exacerbates inequities and hinders school-improvement efforts. Failure to align policies toward a coherent set of practices is also a major impediment, as is a dysfunctional school culture. Implementing effective PD well also requires responsiveness to the needs of educators and learners and to the contexts in which teaching and learning will take place.

Implications for Policy and Practice

Examples of PD that have been successful in raising student achievement can help policymakers and practitioners better understand what quality teacher professional learning looks like. Below are recommended actions for policymakers to support and incentivize the kind of evidence-based PD described here.

• Adopt standards for professional development to guide the design, evaluation, and funding of professional learning provided to educators. These standards might reflect the features of effective professional learning outlined in this report as well as standards for implementation.
• Evaluate and redesign the use of time and school schedules to increase opportunities for professional learning and collaboration, including participation in professional learning communities, peer coaching and observations across classrooms, and collaborative planning.
• Regularly conduct needs assessments using data from staff surveys to identify areas of professional learning most needed and desired by educators. Data from these sources can help ensure that professional learning is not disconnected from practice and supports the areas of knowledge and skills educators want to develop.
• Identify and develop expert teachers as mentors and coaches to support learning in their area(s) of expertise for other educators.
• Integrate professional learning into the Every Student Succeeds Act (ESSA) school improvement initiatives, such as efforts to implement new learning standards, use student data to inform instruction, improve student literacy, increase student access to advanced coursework, and create a positive and inclusive learning environment.
• Provide technology-facilitated opportunities for professional learning and coaching, using funding available under Titles II and IV of ESSA to address the needs of rural communities and provide opportunities for intradistrict and intraschool collaboration.
• Provide flexible funding and continuing education units for learning opportunities that include sustained engagement in collaboration, mentoring, and coaching, as well as institutes, workshops, and seminars.

Well-designed and implemented PD should be considered an essential component of a comprehensive system of teaching and learning that supports students to develop the knowledge, skills, and competencies they need to thrive in the 21st century. To ensure a coherent system that supports teachers across the entire professional continuum, professional learning should link to their experiences in preparation and induction, as well as to teaching standards and evaluation. It should also bridge to leadership opportunities to ensure a comprehensive system focused on the growth and development of teachers.

Related Resources

• Educator Preparation Laboratory (project)
• Teacher Preparation and Professional Learning (topic)
• Developing Effective Principals: What Kind of Learning Matters? (report)
• Educator Learning to Enact the Science of Learning and Development (report)

This report was prepared with the assistance of Danny Espinoza .

Effective Teacher Professional Development by Linda Darling-Hammond, Maria E. Hyler, and Madelyn Gardner is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License .