ArticlePDF Available

Technology Use for Teacher Professional Development in Low- and Middle-Income Countries: A Systematic Review

Authors:

Abstract and Figures

Pre-service education and in-service teacher professional development (collectively termed teacher professional development or TPD here) can play a pivotal role in raising teaching quality and, therefore, learning outcomes for children and young people in low- and middle-income countries (LMICs). However, TPD opportunities in LMICs are limited, unsustained, and often not informed by recent research evidence, and outcomes are mixed. Educational technologies offer potential to enhance formally provided programmes and informal peer-learning forms of TPD. We present the first systematic review of the literature pertaining to technology-mediated TPD for educators of school-aged learners in LMICs, aiming to characterise appropriate and effective uses of technology along with specific constraints operating in those contexts. An in-depth synthesis of 170 studies was undertaken, considering macro-, meso- and micro-level factors during TPD design and implementation in the 40 LMICs represented. Volume of publications increased dramatically over the review period (2008–2020), indicating that the field is rapidly developing. Results largely showed benefits for teachers, but evidence for sustainability, cost-effectiveness or tangible impacts on classroom practice and student outcomes was thin. Promising, locally-contextualised forms of technology-mediated TPD included virtual coaching, social messaging, blended learning, video-stimulated reflection, and use of subject-specific software/applications. We report on the variable effectiveness of programmes and limited attention to marginalised groups. To maximise effectiveness of technology-enhanced TPD, the role of facilitators or expert peers is paramount – yet often glossed over – and the interpersonal dimension of teacher learning must be maintained. Recommendations are made for researchers, policymakers, teachers and teacher educators.
Content may be subject to copyright.
Computers and Education Open 3 (2022) 100080
Available online 17 February 2022
2666-5573/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Technology Use for Teacher Professional Development in Low- and
Middle-Income Countries: A systematic review
Sara Hennessy
a
,
*
, Sophia DAngelo
b
, Nora McIntyre
a
,
#
, Saalim Koomar
c
, Adam Kreimeia
d
,
Lydia Cao
b
, Meaghan Brugha
c
, Asma Zubairi
d
a
EdTech Hub and Faculty of Education, University of Cambridge, 184 Hills Road, Cambridge CB2 8PQ, UK
b
EdTech Hub and Open Development & Education (ODE), 3 York Terrace, Cambridge, CB1 2PR, UK
c
EdTech Hub and Jigsaw Consult, Work.Life Hammersmith, Kings House, 174 Hammersmith Road, London W6 7JP, UK
d
Overseas Development Institute (ODI), 203 Blackfriars Rd, London, SE1 8NJ, UK
ARTICLE INFO
Keywords:
Teacher professional development
teacher training
technology
ICT
low- and middle-income countries (LMICs)
e-learning
pedagogy
peer learning
ABSTRACT
Pre-service education and in-service teacher professional development (collectively termed teacher professional
development or TPD here) can play a pivotal role in raising teaching quality and, therefore, learning outcomes
for children and young people in low- and middle-income countries (LMICs). However, TPD opportunities in
LMICs are limited, unsustained, and often not informed by recent research evidence, and outcomes are mixed.
Educational technologies offer potential to enhance formally provided programmes and informal peer-learning
forms of TPD. We present the rst systematic review of the literature pertaining to technology-mediated TPD
for educators of school-aged learners in LMICs, aiming to characterise appropriate and effective uses of tech-
nology along with specic constraints operating in those contexts.
An in-depth synthesis of 170 studies was undertaken, considering macro-, meso- and micro-level factors during
TPD design and implementation in the 40 LMICs represented. Volume of publications increased dramatically
over the review period (20082020), indicating that the eld is rapidly developing. Results largely showed
benets for teachers, but evidence for sustainability, cost-effectiveness or tangible impacts on classroom practice
and student outcomes was thin. Promising, locally-contextualised forms of technology-mediated TPD included
virtual coaching, social messaging, blended learning, video-stimulated reection, and use of subject-specic
software/applications. We report on the variable effectiveness of programmes and limited attention to mar-
ginalised groups. To maximise effectiveness of technology-enhanced TPD, the role of facilitators or expert peers
is paramount yet often glossed over and the interpersonal dimension of teacher learning must be maintained.
Recommendations are made for researchers, policymakers, teachers and teacher educators.
1. Introduction
Learning outcomes for children and young people in low- and
middle-income countries (LMICs)
1
are well below expected levels [251].
UNESCO Institute for Statistics (UIS) estimates that more than 380
million children worldwide (56% on average and 85% in sub-Saharan
Africa [SSA]) will nish primary school without being able to read or
do basic mathematics [221]. Myriad reasons underlie low educational
attainment, including the quality of teaching, but the all-too-prevalent
decit model is inappropriate [178,212].
Pre-service education and in-service teacher professional develop-
ment (TPD) can be pivotal in supporting teachers to be effective edu-
cators [50,178]. Teacher education interventions are associated with
positive effects on primary school learning outcomes in LMICs [6,51,
129]. However, teachers in LMICs often lack access to in-service TPD
opportunities and, as in high-income countries (HICs), there has been
Abbreviations: DBR, design-based research; EiE, education in emergencies; HIC, high-income country; LMICs, low- and middle-income countries; LIC, low-income
country; MIC, middle-income country; MOOC, Massive Open Online Course; OER, open educational resource; SSA, sub-Saharan Africa; TPD, (pre-service and in-
service) teacher professional development; VLE, virtual learning environment.
* Corresponding author: Faculty of Education, University of Cambridge, 184 Hills Road, Cambridge CB2 8PQ, UK
E-mail address: sch30@cam.ac.uk (S. Hennessy).
#
Present address: Southampton Education School, University of Southampton, University Road, Southampton SO17 1BJ
1
LMICs are dened as those listed in the World Banks country inventory.
Contents lists available at ScienceDirect
Computers and Education Open
journal homepage: www.sciencedirect.com/journal/computers-and-education-open
https://doi.org/10.1016/j.caeo.2022.100080
Received 17 December 2021; Received in revised form 15 February 2022; Accepted 15 February 2022
Computers and Education Open 3 (2022) 100080
2
considerable variation in the efcacy of programmes for both teacher
and student learning [127,201,251].
Increasing the use of technology across education systems offers
potential benets for mediating TPD. Educational technology (EdTech)
2
can facilitate the effective delivery of TPD, especially in remote areas
and for teacher peer support, one-to-one coaching and reection
through exploiting affordances
3
such as two-way communication, au-
diovisual media capability and ease of access to mobile devices.
4
Technology integration brings additional needs to upskill teachers. For
example, learning gains from personalised, adaptive technology use are
greater when an experienced teacher is available to offer contextualised
input and feedback [121]. A systematic review of research on EdTech to
support learners with disabilities in LMICs likewise emphasises the
pressing need for TPD [116]. However, EdTech interventions that focus
only on hardware/software are ineffective compared with those coupled
with other measures to promote pedagogic change or teacher develop-
ment [6,51,129,252]. Lack of appropriate, sustained support for teach-
ers to develop condence in using digital technology before classroom
implementation is a major obstacle to effective technology use in the
classroom [10,36] and to widespread rollout of technology-supported
initiatives in schools [229]. The Covid-19 pandemic and widespread
school closures have rapidly escalated the need to offer teachers various
support options for using technology for remote teaching and learning,
including digital pedagogical strategies and mental wellness support
[33]. However, teachers in LMICs have been consistently asked to
communicate with and offer instruction to learners during the
ongoing pandemic without the appropriate training [234].
Professional learning mediated through technology use can poten-
tially contribute to improving teaching quality and student learning
outcomes. In this review, we critically appraise the evidence in the eld,
using a broad denition of the shorthand term ‘TPDto encompass all
forms of pre- and in-service teacher development that support teaching
and student learning and on-/off-site provision including formal pro-
grammes, mentoring and coaching as well as informal teacher learning,
for instance through online peer communities of practice. We conducted
a comprehensive review of the literature published between 20082020
pertaining to EdTech as a means of delivering initial teacher education
or professional development (‘Tech for TPD) in LMICs. This included
improving subject knowledge and/or classroom pedagogy.
Over the past decade the number of publications on EdTech for TPD
in LMICs has increased signicantly (see Figure 4 and publications such
as [4,178]). However, there are few evidence reviews within the eld,
and those few are often limited in their scope, for instance through
focusing on a specic TPD modality (e.g., MOOCs: [76]; or mobile
instant messaging: [211]), on region (e.g., [228]) or on target skill (e.g.
teachers critical thinking: [204]) or group (e.g. learners with disabil-
ities: [116]). Most reviews related to EdTech for TPD do not focus on
LMICs nor even include many LMICs in the mix with HICs; in some, no
studies from Latin America or Africa are represented [41].
This is the rst systematic review of the literature across the eld of
technology-mediated TPD in LMICs. It makes signicant contributions
by:
taking stock of the research eld and consider its key themes, mes-
sages, strengths, weaknesses and gaps.
identifying promising uses of technology in TPD in LMICs; formats
and outcomes of TPD initiatives, including design principles for
scalable and sustainable design; and the enabling and constraining
factors that impact on effectiveness.
offering rigorous evidence to inform future programming and
research concerning new professional learning initiatives.
The prior research and the focus of the review are outlined in Section
2. The methodology of this review is elaborated in Section 3 and an
overview of the demographic characteristics and research designs in the
170 publications appears in Section 4. The ndings (Section 5) report
the development and efcacy of technology for TPD models across
diverse cultures and contexts. The article concludes by offering practical
recommendations to policy makers, researchers, teachers and teacher
educators (Section 6).
A further, unique contribution to the eld is the full resulting data-
base with detailed thematic coding and quality assessment which is
made openly available (Supplementary material) for researchers and
other stakeholders.
2. Conceptual framework
2.1. Characteristics of effective TPD in LMICs
There is a strong consensus across the eld concerning the charac-
teristics of effective contemporary models of TPD (listed in Table 1).
However, impact tends to focus on teaching practices rather than stu-
dent outcomes, where causality is hard to determine (as Schwille and
Dembele [193] highlight). Moreover, the impact of specic factors
Table 1
Key characteristics of effective TPD
Characteristic Article
sources
Programme content and relevance to teachersneeds
Informed by research evidence concerning effective pedagogy [75,164]
Builds on teachersexisting knowledge, expertise and practices [35,215]
Integrates subject knowledge and pedagogy [54,163]
Participatory and addresses teachersneeds, constraints, interests
and agendas
[5,178];
Engages with difference and ensures linguistic, socio-cultural and
other context-specicity
[220,233]
Recognises that students have diverse learning needs [94,226]
Mode of delivery and support mechanisms
Focuses on imminent practical application, supporting iterative
cycles of trial and renement of new approaches through reective
inquiry and rehearsal within a safe environment
[37,176]
Programme design models the pedagogic approaches being
promoted, taking a learner-centred approach
[73,155]
Emphasises peer dialogue and collaboration within a community of
practice (including coaching)
[94,235]
Opportunities for discussion and critique of alternative approaches [125,241]
School-based and teacher-led [5,32]
Institutional and external support for participation
Alignment with national policy, curriculum standards and
assessment frameworks
[106,228]
Alignment with institutional strategic goals and support from school
leaders, striving for school-wide participation and impact
[107,127]
A conducive culture for professional learning, reection and
feedback based on trust and support
[108]
Dedicated time for professional development and collaborative
inquiry
[39,184]
Programmes are sustained over time through regular sessions and
ongoing scaffolding
[201,219]
5
5
This source synthesised ndings from Cordingley et al., [31], Hall [62],
McCormick et al., [128], Murchan et al., [148], Ofsted [156], Opfer et al., [161],
Pedder et al., [170], and Williamson & Morgan [244].
2
‘EdTechis dened as technologies including hardware, software and
digital content that are either designed or appropriated for educational pur-
poses. We deliberately use a broad denition of EdTech that includes any use of
information and communications technologies (ICT) at any point within the
education system in ministries, schools, communities and homes, including
between individuals and for self-study. Most EdTech relates to digital tech-
nologies but low-tech devices like non-digital radio and television are included
as potentially appropriate for the most marginalised groups such as those in
remote rural areas [53].
3
In this article we conceptualise affordances as the perceived, i.e., uid
rather than xed, properties of a technology that can be exploited for a specic
purpose in a specic context [55].
4
See blog by Patil & Kaye (2021).
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
3
cannot easily be distinguished and evidence for some characteristics in
LMICs is weak [90,201], with much of the research coming from HICs.
Nevertheless, the evidence base on these characteristics is growing in
LMICs.
While these same principles seem to apply across geographies (e.g.,
[63,143]), they require certain considerations and adaptations to attend
to the diverse socio-cultural and economic contexts of LMICs. A scarcity
of teaching and learning resources, limited technology infrastructure
(especially internet and electricity connections), and limited funding for
TPD are common examples (see Section 2.3). A further consideration
arises concerning the intended outcomes of TPD programmes. Teachers
in LMICs often experience inadequate, theory-heavy pre-service educa-
tion, and, as a result, can have low levels of pedagogical or subject
content knowledge (Westbrook et al., 2014). While TPD programmes
focusing exclusively on subject content knowledge tend to have less
effect on student learning [90], pedagogically-focused TPD benets
from practical application to specic subjects, especially at secondary
level. In-service provision is not always linked to classroom applications
though, nor sustained over time ([90,163];). These constraints become
exacerbated in remote regions or areas of extreme poverty, where there
is often an absence of both equitable opportunities for TPD and in-
centives to participate in those that do exist [178].
Further tensions concern the most valuable kinds and levels of TPD
support for teachers in LMICs, including facilitation by advanced
teacher peers or external experts/coaches. The value of leveraging
external expertise in exposing teachers to new ideas is recognised,
enabling groups to avoid becoming overly inward-looking [32] and
challenging local norms and shared values [180]. We argue, though,
that it is important to construe all teachers as professionals capable of
reecting on, critiquing and developing their own practice and recog-
nising their own agency to effect change. This includes adapting and
testing new approaches in their own contexts to solve local problems
[108,192], and identifying and addressing gaps in their own knowledge
and skills [64]. That said, teachers are a heterogeneous group. They vary
in their condence, knowledge and skill levels, along with their work-
loads and time available for engaging in professional learning and
reection. Thus, some may require more support and perhaps more
structured resources than others; whether less structured approaches
can scale up is an open question to be addressed. This thorny issue of
how to offer effective TPD at scale while maintaining sensitivity to the
local context is explored in the next section.
2.2. Coherent TPD systems
The characteristics discussed above are derived from research but
their adoption does not always guarantee positive outcomes. Indeed,
top-down, ‘one-size-ts-allor ‘best practiceprogrammes without built-
in opportunities for local contextualisation are problematic (e.g., [215]).
Thus, the core characteristics outlined above must not be considered
prescriptive, nor universal, but rather adaptive and exible.
Context-specicity is the paramount characteristic that determines the
possibility to engage with and adopt the other important facets we know
bring about effective professional learning and its classroom
application.
Approaching TPD holistically, from a ‘systemsperspective, helps us
understand the people and processes that both inuence and are inu-
enced by the system. A review by Opfer & Pedder [162] proposed three
overlapping systems involved in professional learning. First, the authors
describe the individual teacher as a system that encompasses their
prior experiences, worldviews, beliefs, attitudes and (pedagogical and
content) knowledge. Second, the school-level system includes a col-
lective culture of beliefs, norms and practices. The third system of ac-
tivities within the TPD design consists of the activities and materials
used, and the frequency with which they are employed. This multilevel
lens shows the considerations needed to establish a coherent TPD sys-
tem. Missing from Opfer & Pedders [162] review, however, is a
discussion of the role of technology use in TPD. Nor do the authors
discuss how macroscopic systems (e.g., political environments and
infrastructure) inuence the teacher, school and TPD activities.
The concept of coherencein TPD [112], furthers Opfer & Pedders
[162] systems narrative. Coherence manifests threefold:
1) external coherence at both the national (macro-) level (e.g., with
policy, curriculum standards and evaluation systems) and local
(meso-) level (e.g., with school leadership and resources);
2) internal coherence (e.g., between activities within the TPD pro-
grammes);
3) coherence creation the development of goals that are either pre-
determined or negotiated together with individual teachers (micro-
level).
At the micro level, Lindvall and Ryve suggest that only through
coherence creationare teachers given the requisite agency in making
decisions regarding their professional development. Again, though, they
do not address the role of technology.
Teachers know their own contexts, learners and needs best. Needs
assessments and participatory approaches to TPD are likely to be fruitful
in delivering strong outcomes, especially in the challenging and diverse
contexts of teaching in LMICs (e.g. Anamuah-Mensah et al., 2013).
However, teachers voices are often neglected in the design of and
decision-making around TPD programmes. Related system-level bar-
riers to effective TPD include a potential misalignment among govern-
ment policy statements, institutional cultures and individual
professional responsibility, and varying degrees of autonomy within a
school or for schools themselves [219]. In some schools, teachers may be
restricted to delivering lessons and initiatives planned by their school or
district, constraining the testing and development of new pedagogical
approaches that TPD aims to foster.
Participatory approaches to TPD include culturally relevant,
linguistically accessible and suitably pitched content and they involve
teachers in decisions around what student-centred practice might look
like within their context [21,73]. In essence, TPD must be designed with
teachers, not for them because the structures and cultures within which
teachers work inevitably shape their everyday practices ([18], p. 39).
Thus, co-design is construed as a signicant element of a
well-functioning education system, where teachersvoices are heard
and listened to at the school and wider policy levels (e.g., [82]). This
helps create a reciprocal transfer of knowledge, experience and learning.
However, in both LMICs and HICs, teaching is a low-status profession
[42,188] and teachers lack opportunities to actively participate in
shaping decisions that affect them.
We have blended concepts from Lindvall and Ryve, and Opfer and
Pedder and included a technological lens for this review Figure 1.
illustrates the importance of considering TPD programmes as existing
within a broader system and shows how micro-level factors such as
teachers individual beliefs or backgrounds t into a wider school or
community culture (meso-level). This can then generate a broader
workforce culture in which teachers are situated (macro-level).
This section has suggested how coherence between the different
levels of educational systems can produce effective, scalable and equi-
table TPD. If these levels are all considered, the specic contexts in
which key educational stakeholders operate should complement rather
than constrain one another. The present review emphasises consider-
ation of how the different notions of coherence affect the position of
teachers within tech-supported TPD initiatives in LMICs and what this
implies for associated learning outcomes for both teachers and students.
2.3. The role of technology in the TPD system
Technology has been integrated into TPD models for over two
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
4
decades (see historical overview of publications in Appendix A). A
number of publications explored the role of technology within TPD,
including in LMICs, in the early 2000s (e.g [53,94,228].). These incor-
porated many of the important characteristics, principles and consid-
erations noted in Section 2.1 that remain pertinent for contemporary
education planning and practice. For instance, a 2015 World Bank [250]
guide focused on the use of computers, internet, broadcast media (radio
and television), video and online distance learning [53]. While new
technologies such as tablets and smartphones have risen rapidly in
prominence, the eld of technology-mediated TPD has lagged behind.
This prompts further questions and highlights the importance of iden-
tifying and addressing barriers to adopting effective approaches and
advancing innovation in the eld.
A noticeable and expected trend across publications over the past
two decades has been a shift from theoretical and conceptual guides and
frameworks, to research addressing specic, persistent issues and
identifying effective practices drawing on evidence from TPD initiatives
[106,127,216,235]. For instance, a distinction between publications
focusing on LMICs from the early 2000s and late 2010s is the relative
emphasis placed on the technology itself versus human factors involved
in TPD. Later research focuses more on understanding teachers as in-
dividuals, listening to their needs, promoting professional autonomy
and enabling collaboration. This demonstrates that there have been
signicant yet often subtle changes over time in understanding how
technology can effectively support TPD in LMICs. This highlights the
need for analysis on the extent to which these changes are being
implemented in practice.
The eld studies, like much research in EdTech generally, often take
an uncritical, aspirational approach to the use of technology [194].
Many gloss over the actual added value of technology, or its cost and
reliability implications compared to non-digital modes [74]. These
considerations are vital, since LMICs have limited technological infra-
structure; computers in particular are expensive and school systems
often struggle to integrate the technology into teaching and learning
[252]. TPD programmes reliant on technology availability will likely
cost more than in HICs [68,122]. Costs of software/digital content may
be prohibitive unless these are openly available [183]. Low-tech devices
and resources that run on low bandwidths or do not rely on continuous
power supply may be more reliable and cost-effective [27]. Low-cost
technology such as broadcast media reaches mass audiences in remote
areas
5
(e.g. the Rising on Air programme: [104]), however it can be less
interactive than high-tech options.
Many digital divides exist between (and within) LMICs and HICs; for
instance, limited access to technology results in lower levels of digital
skills in teachers [115,225]. The time needed for teachers to familiarise
themselves with new technologies is thus greater. Basic mobile phones,
TV and radio (or ‘lower techmodalities), which tend to be more ubiq-
uitous in LMICs than ‘higher tech modalities such as tablets and
smartphones may therefore be desirable, albeit under-exploited, op-
tions. For example, access to mobile devices among adults has reached
100% in some African countries (e.g., South Africa, Namibia and
Zimbabwe), although a rural-urban divide persists in most, and women
and girls often have less access than men [105]. Nevertheless, even
when teachers have access to mobile technologies, they may be reluctant
to use them for professional purposes, for reasons including having ac-
cess only to a shared device and the high costs of data [120,134].
6
EdTech has a potential role to play in LMICs when addressing needs
of marginalised communities. Laurillard et al. [105] argue that online
learning at scale can address the issue of equity in teacher education.
Female higher education teachers in Pakistan, for instance, reported that
digital platforms gave them a voice whereas negative gender attitudes
prevented them from engaging in face-to-face discussion with peers
[93]. TPD can also be used to improve teachersability to exploit
technology in adapting to diverse learnersneeds, tackling prevalent
barriers to access to education for marginalised children and young
Figure 1. A coherent tech-supported TPD system
5
Some radio shows or TV broadcasts, for example, may allow students or
teachers to call in for individualised support or instruction.
6
Data costs in many LMICs, especially in SSA, are much higher than in the
UK and Western Europe, for instance, see: https://www.cable.co.uk/mobiles/
worldwide-data-pricing/.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
5
people. There is promising evidence concerning remote or online
learning in fragile and/or conict-affected settings supporting learning
of both displaced teachers and students [106,214]. However, caution is
needed since research indicates that many kinds of technologies are
disproportionately used by the most privileged learners within each
LMIC [25,113,195].
7
Differential access to all technologies along
geographical (urban/rural, district/national/regional) divides is well
documented and was exacerbated during the Covid-19 pandemic [44,
234]. In addition, the digital gender divide may affect female teachers
attitudes towards, and use of, digital technologies [230]. Millions of
teachers, like learners, are further disadvantaged by the fact that most
software and online content is only available in English and Chinese and
not culturally appropriate in many contexts [183].
New challenges arise with the use of technology for TPD in LMICs,
beyond the obvious technical demands. At the micro- or individual
actor-level (as per Figure 1), teachersbeliefs about the nature of
learning determine whether or not they will use technology to imple-
ment learner-centred pedagogies [47]. Teachers attitudes toward
technology have also very commonly been found to inuence their use
of EdTech in the classroom (e.g., [14,134,217]). At the meso or
school/teacher college level, institutional culture inuences the degree
to which technology is adapted and implemented [95], and institutional
leaders play a signicant role in this culture setting [127]. At the macro
or district/national level, the degree to which micro- and meso-level
data are collected and used effectively can inform sustained effective-
ness of programmes [34]. The interaction between different levels of
context must be coordinated in order to enable the coherent system
required for successful implementation.
2.4. Focus of the review
The conceptual framework and gaps in the literature outlined above
led us to the following overarching research question (RQ) for the
review:
How can the use of technology support TPD in LMICs, including
improving subject knowledge and/or classroom pedagogy? (‘Tech
for TPD)
Sub-questions are depicted in Table 2. First, we addressed a set of
descriptive questions in order to (quantitatively) summarise the main
characteristics of the studies reviewed (RQ1). Then we addressed the
main RQ through an in-depth narrative synthesis. The scope included
the modes and outcomes of (in)effective programmes (RQ2). The team
conducting this review is part of EdTech Hub, whose mission addresses
the pressing need to build robust evidence concerning EdTech in-
terventions that improve learning outcomes for the most marginalised
learners in LMICs. Hence RQ3 was framed to focus on equity. Finally,
local contextualisation, sustainability, scalability and cost-effectiveness
(RQ4) are all crucial considerations in any form of teacher learning or
technology use, especially in LMICs [74].
3. Method
An initial systematic mapping review [96] provided a holistic overview
of the whole research eld: use of EdTech related to teacher learning in
LMICs (including teacher learning about technology that supports both
student learning and teaching, planning and assessment practices, as
well as technology for mediating TPD itself). The quantitative outcomes
from that review are presented in a separate technical report [71]. The
mapping review informed the decision to subsequently focus on the
most prominent theme emerging: EdTech use as a medium for TPD. We
conducted an in-depth systematic review of this literature and that is the
focus of this article. The Preferred Reporting Items for Systematic Re-
views and Meta-Analysis (PRISMA) methodology [139] employed
involved two rounds of screening against inclusion criteria followed by
in-depth thematic coding and data extraction. We then appraised the
quality and synthesised the ndings from the research evidence. This
prioritised the higher quality research studies. The overall review pro-
cess of eight steps, from dening the review scope to selecting the nal
papers for analysis, is illustrated in Figure 2.
Three elements of the methodology screening, thematic coding and
quality scoring were carried out by 35 researchers, each handling a
subset of the data. The fourth element data extraction was conducted
by a single team member. In each case, an extensive training and cali-
bration period (lasting several weeks) preceded data processing. Re-
searchers familiarised themselves with the relevant scheme (inclusion/
exclusion criteria in screening, coding scheme, quality scoring frame-
work, data extraction variables) and contributed to its development
through pilot testing and discussion. In this stage, blind parallel pro-
cessing of the same papers by team members facilitated close alignment
between them and iteratively eliminated any ambiguities within the
schemes. During this stage and subsequently, all questionable cases were
resolved through discussion among the team and consultation of the
lead author where appropriate. Ongoing calibration between team
members was ensured through regular spot checks involving blind
screening, coding or scoring of the same records, before moving on to
the next round of the process. Finally, four researchers shared leadership
of the narrative synthesis, each focusing on one of the main themes
emerging (Section 5).
Table 2
Research questions of this study
# Research Question and sub-questions
RQ1 What are the key characteristics of the emerging ‘tech for TPDstudies
in LMICs?
When and where (country/region) was the research published?
Which technology devices and resources were used?
Who designed the TPD programmes? Were designs participatory?
What was the scope of the samples researched? Which teacher populations
did the TPD initiatives reach? What education levels and subjects did they
teach?
Which outcomes have been promoted?
RQ2 Which forms of technology-mediated TPD are effective for teacher
learning?
Which modalities and uses of technology are effective in promoting teacher
learning?
How can technology support coaches/facilitators or virtual coaching of
teachers?
How can technology support peer communities of practice?
How can technology support teacher reection?
How can technology be used to improve subject knowledge and subject
pedagogy?
Can technology support the development of new resources in low-resourced
contexts? If so, how?
How can technology be used to offer exible learning environments with the
appropriate support and levels of structure to meet teachersneeds?
RQ3 Do any technology-mediated TPD initiatives help foster more equitable
education systems?
If so, how do they help reach teachers in remote/rural contexts?
How do they support teachersdevelopment in the language of instruction?
In what ways do they support marginalised teachers (e.g., refugee teachers,
teachers with disabilities, etc.)?
How do they support teachers in addressing the needs of marginalised
learners?
RQ4 In what ways are TPD initiatives systemic in their approach,
considering macro-, meso-, micro- and cross-cutting factors during
design and implementation?
Are technology-mediated TPD initiatives targeted to teachersneeds and
locally and culturally contextualised? Which multilevel factors are pertinent?
Is there evidence for which modalities and uses are cost-effective in LMICs?
Which technology-mediated TPD initiatives are sustainable and scalable?
7
Increasingly stratied education systems mean that both learners and
teachers in elite private/international schools are differentially able to take
advantage of resources, including access to technology.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
6
Figure 2. PRISMA diagram
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
7
3.1. Literature search
3.1.1. Search strategy
The focus of our search was the intersection of the three elds
(depicted in Figure 3). A provisional a priori set of associated key terms
was generated by the team based on knowledge and perusal of titles,
abstracts and keywords of key literature in the eld; it was trialled and
iteratively rened through a series of preliminary searches using
ProQuest.
The resulting terms for the primary search are listed in Table 3. These
were applied to locate academic and grey literature from four electronic
databases: ProQuest Education, British Education Index, Applied Social
Sciences Index & Abstracts (ASSIA) and the World Bank. Supplementary
searches comprised backward-forward searching (using Google Scholar,
based on seeking papers by 8 key authors in the eld) and a snowballing
technique to identify references from seminal literature reviews in the
eld. These searches collectively yielded 204 papers, of which 10 were
independent, met the inclusion criteria and were included in the wider
systematic mapping review. Of these, 9 focused on ‘tech for TPD and
thus were ultimately included in the systematic literature review re-
ported in this article. We also contacted 11 experts in the eld, largely
nominated by the rst author, and 7 responded with a total of 75 rec-
ommended sources, of which 15 were independent, met the inclusion
criteria and were included in the mapping review. Of these, 13 were
ultimately relevant for the systematic review. (Note: both the expert
referrals and the supplementary searches included many relevant papers
that duplicated the initial search outcomes and were therefore
discounted).
3.1.2. Eligibility screening
In correspondence with this reviews focus, our eligibility criteria
comprised three thematic foci.
Theme 1: Studies were included if they collected data from at least
one low- and/or middle-income country listed in the World Banks
[252] country inventory. The search thereby spanned a range of 136
countries in SSA, South Asia, Latin America and the Caribbean,
Middle East and North Africa, East Asia and the Pacic, Europe and
Central Asia (listed in Appendix B).
Theme 2: Studies were included if they investigated teacher learning
i.e., teacher education, teacher professional development, men-
toring or coaching, or peer learning initiatives linked to technology
use. This included situational and contextual factors inuencing
teacher learning through or for using technology, skills and sub-
stantive knowledge, and technology to motivate teachers in their
own learning. Excluded were studies that only reported (without any
link to some form of teacher education/learning) teachersattitudes,
beliefs or perceptions of technology; their self-efcacy, digital skills,
approach to technology use, and intention or readiness to use tech-
nology; or snapshot assessments of knowledge about how to use
technology.
Theme 3: Technology used for teacher development/learning (e.g.,
devices such as computers, video, tablets, or phones, and/or digital
media and software resources). Low-tech devices like non-digital
radio and television were also included since these are often the
only available technologies in remote rural areas. Excluded was
technology for educational management (e.g., data systems, regis-
tration software, emails to parents).
Additional eligibility criteria included:
Literature published in English (20082020, with a cut-off date of 31
July 2020);
Empirical investigations, including primary and secondary data;
excluding reviews and papers that are primarily narrative or theo-
retical in nature;
Peer-reviewed academic journal articles, books and book chapters,
PhD theses, peer-reviewed conference papers, grey literature
(excluding blogs and very short brieng reports);
Focus on pre-service and in-service teachers of students aged 318,
including teaching assistants but excluding school leaders when
researched alone without an accompanying focus teachers;
Formal taught programmes, workshops or informal education/
learning taking place on- or off-site (e.g., in government/private
schools, teacher training colleges, district/community centres),
including through peer communication and support (e.g., via online
communities of practice).
In Stage 1, titles, abstracts and keywords were screened by three
authors to exclude irrelevant texts. In Stage 2, full texts were read by
three authors to exclude further records. Duplicate and non-independent
records were eliminated. Journal articles were prioritised over other
sources such as book chapters and conference papers reporting the same
study. Eligibility screening was carried out in an open-source, dedicated,
Figure 3. Intersecting target search areas
Table 3
Search terms used in the review
Theme
no.
Theme
name
Search terms
17
1 LMIC LMIC OR low middle income countr* OR sub*saharan
africa OR latin america
OR
[Actual LMICs fully listed]
2 TPD teacher development OR teacher training OR teacher
education OR teacher learning OR teacher professional
learning OR
[teacher OR educator OR classroom practitioner OR
school OR instructor] AND [professional development OR
in*service training OR pre*service training OR in*service
education OR pre*service education OR coach* OR
mentor*]
3 Tech technolog* OR digital OR device OR software OR
hardware OR phone OR ICT OR comput* OR video OR
radio OR TV OR televis* OR laptop OR tablet OR learner
management system OR LMS OR virtual learning
environment OR VLE OR e*learning OR blended learning
OR online learning OR mobile learning OR social
network* OR messenger OR messag* OR SMS OR MOOC
OR social media OR professional learning network OR
remote learning OR distance learning
17
The use of an asterisk (*) denotes a wild character used to locate all sin-
gular/plural forms of a term.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
8
systematic review platform, ColandR.
3.2. Data synthesis
Data extraction and thematic categorisations were conducted in
ColandR.
3.2.1. Data extraction
Data extraction was conducted by one author after iterative piloting
testing and renement of the framework by three authors. There is a
wide diversity of settings within reviews of research in international
development, and consequent challenges in determining relevance and
generalisability of ndings across a range of different populations and
‘real world contexts [236]. Thus, we paid particular attention to
describing contextual characteristics in detail.
The nal terms included: publication type, TPD audience, TPD
setting (type of institution, country), sampling strategy, sample size,
clustering basis if applicable (e.g., teaching experience), teacher age and
gender, study design, qualitative data sources, quantitative data sources
and data analytic techniques. A brief summary of ndings from each
study was formulated at the same time. See Appendix C for details.
3.2.2. Thematic coding
Thematic coding of the studies consisted of the broad application of
established qualitative approaches to textual coding [15] through
careful reading and annotation of full texts by three authors. The sub-
stantive expertise among the authors was harnessed to develop a draft, a
priori, thematic framework with high potential for applicability to the
studies. The framework was then tested on increasing ‘sample sizes of
the studies. During the rst iteration, ten studies were coded using this
framework, which was amended and augmented before being applied to
update the same studies and ten more. Between each new cycle, po-
tential changes and ‘emerging codeswere logged for whole-team dis-
cussion and agreement. This iterative framework development process
took place over four cycles. Coding of all studies was then nalised
during the fth cycle (producing the framework detailed in Appendix D)
before exporting thematic data for the whole database of studies.
The framework included: TPD audience, Education Level, TPD
context, Subject, Tech devices, Tech resources, Tech for communication,
TPD modes and methods, TPD design, TPD peer support, TPD outcomes,
Factors at system level, Factors at local level, Marginalised learners.
Cross-tabulations were carried out to explore key relationships of
interest.
3.3. Quality assessment
Studies underwent quality assessment using a criteria checklist
iteratively developed with colleagues across our wider research pro-
gramme (EdTech Hub). This drew in particular on the Building Evidence
in Education (BE
2
) [77] and Mixed Methods Appraisal Tool (MMAT)
[78] frameworks, in consultation with the Research Quality Plus
framework for international development research MacLean & Sen
[119]. As a result, six dimensions were considered and iteratively
rened until a nal version (see Appendix E) was reached and system-
atically applied across the dataset by ve researchers. Dimensions are
summarised in Table 4. One of the dimensions (research design) was
double-weighted due to its particular importance in shaping the aca-
demic rigour of the evidence. In light of the discussion in Section 2, our
quality framework includes the dimension of ‘methodological sensitivity
to culture and ethics. Scores were awarded up to a total of 21 and
banded into four rating categories (1 =Low, 2 =Low-medium,
3 =Medium-high, 4 =High) for the purpose of synthesis. Appendix E
includes a summary of the outcomes of the quality scoring (mean scores
on each dimension and banding distribution) for all articles; detailed
scores for each record are presented in the literature database in the
Supplementary Material.
The quality scoring process included a seventh dimension rating
relevance of the content to the reviews scope and focus on technology
use for teacher learning in LMICs at three levels (1 =low, 2 =medium,
3 =high). This judgement determined how central or peripheral the
study was to our concerns and how much useful information was pro-
vided (see Appendix E for details and outcomes of the relevance ratings).
Low proportions of the 170 records were low quality and/or low
relevance.
Each record was then given an overall rating from 26 calculated by
adding its numerical quality and relevance ratings together; this rough
measure was used simply for our purpose of sequencing the synthesis
through prioritising the most relevant, highest quality papers. In order
to make the challenging task of synthesising across 170 sources
manageable in writing the article, we left aside those rated 23
(although these were still subjected to thematic coding and data
extraction). We started at the other end of the spectrum with the high
relevance plus high-quality records (rated 6), adding in ndings from
the other literature (rated 45) in descending order until saturation
point was reached for themes.
Note that although the quality assessment framework is pioneering
in its holistic nature, a limitation of this tool is that shorter publications,
such as conference papers, unfortunately score lower than extended
reports such as journal publications. We recognise that some sources
designated ‘low qualityor ‘low relevancemay provide important in-
sights. Thus, all of the remaining publications (rated 2-4) were skim read
to ascertain any added value, and a few further points and citations were
added.
3.4. Methodological limitations
The review focused only on publications in English. Systematic re-
views commonly appear to favour publications from English-speaking
researchers in HICs [3]. We experienced difculties in gaining access
to grey literature: it is unclear whether it did not exist or was publicly
unavailable. However, publication of reports to sponsors is often pro-
hibited. It is also possible that some TPD programmes initiated and led
by nongovernmental organisations (NGOs), local partners, and schools
were not published.
The study only reviewed published works and thus could not
incorporate teachersvoices, other than indirectly through reporting on
the small body of research that offered models for co-design of TPD. It
would be useful in the future to solicit feedback and input directly from
practitioners in LMICs. The cut-off date for literature screening meant
that the review also does not cover published works pertaining to the
Table 4
Quality scoring dimensions
No. Description
1 Research conceptualisation, including whether the research report took a
critical approach to the technology use and addressed cultural
contextualisation.
2 Contextual components in the research methodology, including considerations
regarding the sampling strategy as well as reported participant and setting
details.
3 Research design, assessed using the MMAT [78], including qualitative studies,
RCTs, non-randomised studies, quantitative descriptive studies and
mixed-methods studies.
4 Methodological bias; including whether study limitations and researchersown
positioning were considered and safeguarded against.
5 Methodological sensitivity to culture and ethics, including whether data collection
and analyses were culturally appropriate and ethically sound.
6 Claims and conclusions, including whether these were linked sufciently with
prior evidence and evidence emerging from the study.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
9
Covid-19 context that came into being in Spring 2020. Many of the
principles underpinning distance/blended learning models of TPD will
still apply, however. Finally, the large number of records returned meant
that it was not feasible to read all 170 in the same depth; this was
mitigated through closest examination of the most relevant and highest
quality studies (see 3.3).
4. Results characteristics of the research (RQ1)
This section answers RQ1 by outlining the key characteristics of
emerging ‘tech for TPD studies. Information concerning further
extracted variables and codes is available in Appendix F: sampling
strategy and size, teacher age and experience, study design, research
methods, methodologies, data sources, methods of analysis, full
geographic distribution of studies by country, tech devices and
communication modes (by context and region), forms of peer support,
TPD outcomes (by region), and local-/system-level factors inuencing
tech-mediated TPD. The database of 170 sources, complete with coding,
quality scoring and data extraction results, is available in the online
supplementary material here. Readers may wish to use it to conduct
their own targeted analyses.
4.1. Publication numbers and originators
The ndings in this section offer highlights derived from the data
extraction described in 3.2.1, with fuller outcomes tabulated in Ap-
pendix C.
4.1.1. Number of studies published by year
This systematic review found that from 2012 onwards the number of
studies on tech for TPD in LMICs has increased (Figure 4). The gures
conrm that the choice of the time period captured the eld, which is at
an early stage compared to other work on TPD and on uses of EdTech.
4.1.2. Publication types
Figure 5 shows that an overwhelming majority of the studies
emerging (83%) were published via academic journals although journal
quality varies signicantly. While our search criteria included grey
literature, non-academic publications (e.g. technical reports) made up a
surprisingly small proportion (6%): see Methodological limitations
above.
4.2. Characteristics of the TPD studies
The ndings here derived largely from the thematic coding exercise
described in 3.2.2; highlights are presented, while frequencies (f) for
each individual code across the sample are shown in Appendix D. It is
important to note that frequencies will in most cases exceed the number
of studies found for this systematic review since a single variable may
contain more than one code. Thus, the percentages displayed in the
gures depict the proportions of each code out of the total numbers
of codes applied within that category. For example, a single study
could have used multiple sampling strategies (e.g., purposive and
volunteer); as a result, it might have more than one code under the
variable of sampling strategy. The gure used would then be the per-
centage of all instances of sampling strategy that were coded as purpo-
sive, rather than the percentage of studies. On the other hand, n
indicates the number of studies where this is more relevant.
4.2.1. Geographic distribution of research in LMICs
It was notable that 96 out of 136 (71%) LMICs had no studies
Figure 6. breaks this down by region.
Aside from South Asia, all other regions with LMICs had no studies
conducted in the majority of their countries. While more than two-thirds
of countries in SSA had no research, over one-third of the 40 LMICs with
a study were situated within the SSA region.
8
The criteria applied for
this systematic review included only studies in the English language.
This may explain why no studies emerged for the majority of Franco-
phone, Lusophone, Spanish or Arabic-speaking LMICs. The ndings may
also be inuenced by inequitable availability of research funding and
access to publications across countries [3].
Across the 40 countries where a study was based, the majority (28)
had between 1 and 4 studies (Figure 7). See Table 1 in Appendix F for
details on the distribution of studies across the 40 countries. The review
identied just 5 countries which had 10 or more studies China, India,
Kenya, South Africa and Turkey. All of these are classied as MICs and
sit within the top 30 countries globally as far as population size is con-
cerned (Figure 8). Despite the high levels of research outputs for these
ve countries, a question arises as to how far-reaching these studies are
when considering research produced per capita. This is particularly
relevant when considering marginalised groups, and the extent to which
the studies address the high levels of inequality in educational access
and learning outcomes in some of these countries.
4.2.2. Technological devices and resources
Computers were the most common among all mentions of devices
included in studies (Figure 9), followed by smart (internet-enabled)
devices (e.g., smartphones or tablets). Computers (laptops/desktops)
were particularly prevalent in informal education settings (75% men-
tions). Smart devices which have greater reach were at least as
Figure 4. Number of studies published per year: Raw and cumulative fre-
quencies
Note: The tailing off of studies published in 2020 is attributable to this review
only including studies up until July 2020.
Figure 5. Distribution of publication types
8
This is partly due to SSA accounting for the largest number (over one-third)
of global LMICs. Conversely, the higher proportion of LMICs in South Asia with
a study is in part due to the regions relatively small number of countries.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
10
popular as computers in remote/rural areas and community schools,
though (see details, plus regional breakdowns, in Appendix F). More
than half (54%) of the studies that mentioned computers used them
alone. Research featuring all other types of devices (interactive white-
boards, projectors, printers, storage disks, etc.) used them in combina-
tion with other forms of technology.
Studies tended to be concentrated around certain tech resources
(Figure 10). In particular, video resources, web resources and software
app resources collectively comprised 175 out of 294 (60%) references to
tech resources. Other tech resources (e.g., audio resources, Open
Educational Resources [OERs], digital presentations) were frequently
mentioned in combination.
4.2.3. Educational level and subject
A total of 45 studies did not specify the educational level Figure 11.
shows that primary and secondary education levels accounted for
approximately 90% (f=68 and 77 respectively) of mentions. Early
childhood education and vocational education were much rarer foci (4%
and 6% of mentions respectively).
Apart from the 56 studies that did not specify the subject, we found
that the use of technology for teacher learning appears concentrated
around a few subjects. As Figure 12 indicates, second language learning
Figure 6. Share of LMICs by region with and without a study
Figure 7. Numbers of studies by LMIC
Key: No studies, 14 studies, 59 studies, 1014 studies, 15 or more studies.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
11
(SLL) (f=45) and mathematics (f=43) made up over half of all mentions
of subjects. Literacy appeared as a low proportion of total studies
overall, especially in secondary education. Focus subject areas varied by
region Figure 13. shows that more research in SSA than other regions
focused on literacy (often in English as the medium of instruction, often
called EAL or English as an additional language). EAL/SLL was the main
focus of studies in East Asia and the Pacic, Middle East and North Af-
rica, and South Asia. The large concentration of TPD studies on EAL/SLL
in South Asia was related to Bangladesh (f=8), where much research
focused on the English in Action programme [199].
4.2.4. TPD stakeholders
4.2.4.1. TPD designers. In keeping with the concentration of research in
journals (4.1.2), an overwhelming majority of TPD programmes in
LMICs were designed by academic researchers (78%) Figure 14. shows
that participation of local partners (including local community, orga-
nisations and teachers), NGO research groups and governments was low.
School-initiated TPD programmes were also scarce. Accordingly,
participatory methodologies such as design-based research (4%) or ac-
tion research (2%) were rarely used, despite the importance of teachers
input to TPD outlined in Section 2.2 (see Appendix F for further
discussion).
4.2.4.2. Marginalisation. Only 21 of the 170 studies reviewed distin-
guished the characteristics of target students. These references (f=26)
were mainly in the context of socio-economic status (f=11), followed by
learners with disabilities (f=4), orphaned children (f=2), refugee (f=2),
and indigenous community (f=1). A disparate group of studies also
focused on girls, children from remote communities, ethnic minorities,
Figure 8. Countries with the largest numbers of studies
Figure 9. Frequencies of tech device mentions
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
12
children living in conict situations or orphans (total f=6). The numbers
of studies focusing on specic marginalised teacher groups did not reach
the threshold (3 or more) to be included within the coding process.
Therefore, they are not featured in nal frequencies (although examples
appear in Section 5.2.4.1).
4.2.4.3. TPD audience. In-service teachers formed the main focus of
studies (f=105). There were notably fewer studies with pre-service
teachers (f=74). Comparatively few studies focused on unqualied in-
service teachers (f=5) or on teacher educators (f=14). Yet, 25% of in-
service primary teachers in 2019 were estimated to be unqualied in
LICs and 16% for MICs (UNESCO Institute for Statistics [[222], 2019].
The ndings highlight a need for more studies specically relating to
unqualied teachers.
4.2.5. TPD settings
Government schools (f=63) and teacher colleges (f=80) emerged as
the dominant research settings. At the other extreme, there was a dearth
of research focusing on community schools (f=2) or informal contexts
(f=4). From an equity perspective, it is important to note that student
populations in such contexts are likely to come from marginalised
groups.
4.2.6. TPD outcomes
While no single outcome emerged as a majority focus, ‘changed
classroom practices, ‘teacher tech skills/awareness,‘changed peda-
gogical knowledgeand ‘teacher motivation were the outcomes most
commonly addressed (see Figure 15). Considering the two TPD out-
comes generally deemed most desirable [60], ‘changed teacher prac-
tices(f=47) appeared more than twice as often as ‘student outcomes
(f=21). However, together they made up only 17% of the total number
Figure 10. Tech resources used in studies
Figure 11. Educational level of studies
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
13
(f=392) of mentions of TPD outcomes. There were also notable regional
differences (see Appendix F for more detail).
In terms of assessing TPD outcomes, most studies reviewed relied on
self-reporting to measure changes in teacher knowledge and practice.
Self-reporting methods made up 62% of qualitative data (including
interview, focus group, open-ended survey, reection) and 63% of
quantitative data (including questionnaires comprising multiple-choice
questions and rating scales). See Appendix F for details on methods of
data collection. Self-reporting has well-known limitations in terms of
validity. For example, questionnaires used by one reviewed study [56]
demonstrated teachersimproved feelings of self-efcacy towards using
mobile phones as pedagogical tools without linking to actual changes in
classroom practices. 45% (27/60) of the studies that targeted changes in
classroom practice as a TPD outcome were assessed through researcher
observations.
5. Results synthesis of literature on technology use for TPD
Section 5 presents the outcomes of the detailed narrative synthesis
carried out in order to identify, compare and interpret the main ndings
and methodological trends emerging across the 170 studies in the sys-
tematic review. Three main themes emerged: modes of TPD (5.1),
Figure 12. Subject distribution of studies
Figure 13. Subject distribution by region
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
14
equitable systems (5.2) and multi-level factors (5.3). Throughout, we
use Exemplars (in boxes) to spotlight promising tech for TPD practices
that build on these trends in LMICs. These models were chosen because
they meet at least one of three criteria: (i) the model innovatively le-
verages and combines technology uses with TPD programme designs
that are appropriate for LMIC contexts; (ii) the TPD programme led to
changes in teachersclassroom practices or studentslearning outcomes;
(iii) efcacy of the programme was found to be sustainable, scalable or
cost-effective. Although not representative, these models serve as aspi-
rational case studies for researchers, practitioners and policymakers. We
recognise, however, that due to myriad contextual and interdependent
factors affecting TPD rarely will the exact design or structure result in
the same outcomes. It is notable that three of the six initiatives chosen
are (unintentionally) based in Kenya. This partly reects the high fre-
quency of studies in the country, but also that high-quality studies were
conducted there, indicating a strong research landscape to build on and
learn from.
The sections below summarise the headline ndings. We re-address
briey but do not elaborate upon the various contextual factors dis-
cussed in Section 2 when describing the ‘tech for TPDmodels. Instead,
Section 5.3 summarises these contextual ‘multi-level factors inuencing
tech for TPD(RQ4). Again, we recognise the complexities of the TPD
system (see Figure 1). For example, take-up varies greatly across con-
texts and between individual teachers (as with any educational inter-
vention). Likewise, logistical issues (e.g. timing, transportation, costs)
shape whether or not a TPD model is implemented effectively. We also
recognise that the technology itself may be used as a modality, and thus
agency is dependent upon the TPD facilitator or coach who leverages
that modality for teacher learning. Yet, the critical role of the facilitator
or teacher educator and how tech can support it is a highly under-
researched area. Lastly, few studies took a critical approach to the
EdTech used in interventions, rarely detailing any negative or harmful
effects (e.g., safeguarding risks when engaging with online content). It is
important that these risks are documented and understood so that in-
terventions can provide a balanced assessment of the threats and op-
portunities that the use of EdTech presents.
5.1. Effective modes of technology-mediated TPD (RQ2)
The following sections examine the diverse forms of technology-
mediated TPD that have been found to be effective for teacher
learning. It examines variations of TPD models that serve ve common
purposes emerging from the synthesis: to support virtual coaching and
coaches, to foster remote communities of practice, to develop teachers
reective skills, to improve teachers pedagogical subject content
knowledge, and to provide exible learning environments.
5.1.1. Using technology for virtual coaching
5.1.1.1. Virtual coaching and messaging for teachers. Pedagogical
coaching, which typically involves observing, modelling, supporting,
evaluating and offering focused, personalised feedback on teachers
practices iteratively over time, is associated with student learning gains
as well as changes in teacherspractices. A meta-analysis by Kraft et al.,
[102] found similar outcomes for both virtual and in-person coaching.
TPD models that used virtual coaching especially through video
conferencing (f=13) were quite common in the literature reviewed.
There is a promising, growing body of evidence regarding their effec-
tiveness. For instance, in a key study in South Africa providing lessons
plans and other resources on tablets across 180 schools in low-income
communities, a virtual coaching model proved cheaper and as effec-
tive after 1 year as on-site coaching in improving teachersinstructional
practices and learners prociency in EAL [100]. Coaching included
fortnightly follow-up with individual teachers on TPD focus areas,
weekly group messages for motivational support, and sending video
clips relating to teaching areas that teachers found particularly difcult.
However, a recent follow-up study of this programme by [29], published
after the literature review period, indicates that trusting relationships
and lesson observations are harder to secure without face-to-face
interaction. Blended approaches with some in-person visits ensure that
the social aspect of peer support is not lost and they appear to be more
effective
9
although, like leveraging of video-recorded observations, they
Figure 14. Distribution of TPD designers
9
Mary Burns, personal communication regarding an unpublished comparison
study in Indonesia.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
15
are under-researched. (See also Section 5.3.4 on cost-effectiveness and
sustainability of virtual coaching.)
Given the ubiquity of mobile devices in LMICs, Short Message Ser-
vice (SMS) text messages (f=9) in particular emerged as a low-cost,
quick, accessible and familiar form of EdTech to support, prompt or
reinforce new teacher practices [202]. Text messages may best be used
to offer constant, timely reminders or nudges concerning instructional
strategies during or following TPD ([80], Exemplar 1), to hold teachers
accountable for implementing planned classroom activities [152], or to
connect teachers from rural areas (ibid.). Messaging also allows two-way
multimodal communication, including via emoticons, photographs,
videos and les [211]. However, it is only a small part of coaching and
cannot facilitate the observational and feedback elements. It may also be
less effective in addressing complex social issues such as gender norms
(e.g., [28]).
5.1.1.2. Technology support for pedagogical leaders. Tech-enhanced TPD
may also be used to support coaches, mentors and facilitators, for
example by providing more structured observation tools or feedback
guides. In Uganda, Tangerine:Tutor (now ‘Coach) improved the quality
of in-person coaching through coaches using guided observation pro-
tocols [185]. Quantity and coverage of coaching were increased through
receiving automatically generated, relevant feedback after lesson ob-
servations and case management summaries via a central web-accessible
dashboard. However, coaches visited schools less often compared to
previous paper-and-pencil methods, possibly because the digital class-
room observation tool required more data input and increased the time
spent per school. Nevertheless, the tablets and software resources made
the work easier and increased coachescommitment and accountability
(Section 5 discusses accountability further).
A second study with highly promising results is Bruns et al.s [22]
randomised evaluation of a programme in Brazil that provided sec-
ondary schools with classroom observation feedback and virtual
coaching. External expert coaching of the schoolspedagogical co-
ordinators via Skype cost $2.40 per student, proving cost-effective in
raising teachersclassroom effectiveness and producing signicant stu-
dent learning gains in mathematics and Portuguese.
5.1.2. Using technology to foster communities of practice
5.1.2.1. Teacher-led uses of social media to share practice. Social media
(f=38) such as Facebook or WhatsApp may be an effective approach
to TPD due to its versatile nature [61,87]. Research suggests that indi-
vidual factors tend to play a role in the effectiveness of such TPD models,
including how teachers access social media from their mobile devices for
professional learning [8]. Social media is generally an easily accessible
and hence motivating platform. In Ghana, for example, pre-service
teachersuse of smartphone applications such as WhatsApp to share
resources supported their professional development through both
formal college courses and informal off-campus group interaction [210].
Studies in Nigeria and Zimbabwe found that pre-service teachers were
intrinsically motivated to use their mobile phones to access social media
for educational purposes, despite challenges related to ICT infrastruc-
ture and internet costs [149,242].
Teachersknowledge, (ICT) skill sets, and overall dispositions shape
their engagement with social media platforms [206,238,243]. Moodley
[141] found that teachersattitudes towards using an online social
media network (e.g. WhatsApp) to support a virtual community of
practice are dependent on their awareness of the context within which
the community exists and the willingness of the participants to accept
differing views and opinions. Because active participation in online
discussions can provide benets Bett & Makewa [16], incentives to
participate in online TPD (e.g., through promotion opportunities) may
be more effective for teachers who lack professional autonomy or ICT
skills [243]. However, a review of social media use raises the risk of
teacher-led models spreading inaccurate content when communication
channels are not monitored by expert teachers [124].
5.1.2.2. Uses of social media in formal programmes. Social media can be
benecial in blended learning pre-service TPD models for peer assess-
ment of teaching materials. Facebook groups were used this way in a
series of studies (predominantly in Turkey) for pre-service teachers [38,
81]. Pre-service teachers engagement in face-to-face teaching sessions
increased [38], and closed Facebook groups enabled peer evaluation to
extend beyond classroom hours [49].
University lecturers in Malaysia have used social media platforms to
facilitate remote lectures and supervision of pre-service teachersonline
teaching [138]. Social media is also used to incentivise participation in
TPD programmes. In-service teachers in Indonesia felt more incentivised
to engage with an online platform when social media was paired with
in-person TPD and monitored by a teacher educator; however, text
messaging (and phone calls) did not improve participation in online TPD
that provided access to resources on pedagogy and classroom inquiry
[243].
Research also suggests that different social media platforms may
have different affordances. Sun et al., [207] found that both online
discussion forums and a mobile instant-messaging app facilitated
collaborative learning among pre-service teachers. However, the online
discussion forum resulted in more communication aimed at knowledge
construction, while using WhatsApp resulted in more social interactions.
Likewise, a South African study concluded that using WhatsApp meets
pre-service teachersneeds and supports collaboration, but does not
adequately develop digital literacy skills [117]. Yahoo Group was used
asynchronously (i.e., responses to posts were not immediate) by
in-service teachers in Iran both to construct knowledge and provide peer
support [150]. Teachers discussed diverse topics including the use of
ICT, the creation of teaching and learning materials, and classroom
management. Again, impacts on practice are unknown (see Section
5.1.4).
Six studies used social media and other online platforms (wikis, open
Exemplar 1
The Health and Literacy Intervention (HALI) project in Kenya
An intervention in Kenya employed training workshops, semi-scripted lesson plans and weekly SMS text-message support for teachers providing
brief instructional tips and motivation to implement lesson plans. Despite a relatively low level of in-person support, it signicantly impacted
classroom practices. It raised primary childrens literacy outcomes after two years, and reduced school dropout rates [80]. Teachers considered
the weekly text messages a good source of new teaching ideas and the bidirectional SMS model created a supportive community. The effect of
using technology could not easily be isolated, however, as is common. (See another study of the HALI project in Exemplar 7.)
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
16
and distance learning [ODL], virtual learning environments [VLEs]) to
connect teachers from LMICs to peers in HICs [17,23,132,190,205,232].
Despite evidence that teachers prefer face-to-face meetings [23], sharing
practices using multimedia (including videos and images) was effective
in mitigating interpersonal distance, which can cause misunderstanding
and disengagement [190]. Such transnational remote-learning com-
munities can allow peers from different cultural settings to identify
common concerns and deconstruct their assumptions or unexamined
beliefs, leading to more profound self-knowledge [205]. In other words,
a deeper, more critical reection is possible when tech supports TPD to
foster reection across geographical and cultural borders.
5.1.3. Using technology to support teacher self- and peer-reection
Video-based, self-reective TPD is common (f=73 studies) and has
proven especially useful for teachers to see themselves teach and eval-
uate their own practice (e.g., [85]). It can also be used to assess peers or
view expert teachers modelling effective pedagogy. This is especially
helpful for pre-service teachers in contexts where quality teaching may
be elusive [165]. Reecting on videos can lead to changes in teachers
beliefs, subject knowledge, and/or students perceptions of teachers
practices (Ragatz and Sugiarti, 2015). Several studies examined how
videos were effective when paired with structured opportunities for
reection, questioning and discussion [110,165,209]. Additional
artefacts such as lesson plans or journals were used to facilitate
reection [59]. One TPD model provided lecture notes at the beginning
and end of the video, making clear links to theory to support teachers
knowledge of the scientic approach [209]. Conversely, another study
was found to increase teacherstheoretical knowledge (on inclusive
education); however, the videos did not provide examples of practical
application of the theory [98]. Video-based TPD that is linked to
imminent classroom testing of new approaches and reection with peers
may lead to more sustainable change in teacherspedagogy (Exemplar
2).
The type of video resource and the design of the TPD model play a
role in shaping outcomes. Research in Cambodia found that teachers
working in small groups and watching videos of expert teachers who
modelled effective pedagogy were most effective [114]. Without ob-
serversphysical classroom presence, teachers authentic pedagogical
skills are better exposed and the information gathered and feedback
provided to teachers are likely to be more credible [110]. Nevertheless,
major ethical and technical considerations arise with teacher surveil-
lance of this kind. Factors that shape the effectiveness of videos in TPD
models include an institutional culture and leadership supportive of peer
learning, and the costs of tech infrastructure and maintenance (e.g.,
Figure 15. TPD outcomes
Exemplar 2
OER4Schools in Zambia
In the multimedia OER4Schools programme in Zambia, unique videos of interactive teaching by local teachers offered an external stimulus for
teacher discussion, reection with peers, and inquiry. Learning was guided by built-in prompts for both teachers and facilitators, and the
materials explicitly linked theory to practice. Teachers collaboratively designed and trialled new pedagogical strategies. After one school year,
teachers adapted to learnersknowledge levels, and used more practical and group work. In turn, pupils built deeper understandings of subject
matter, collaborated and used digital technologies for problem solving [68,73]. The extensive multimedia materials helped the programme
become self-sustaining after the intervention (see 5.3.4.2).
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
17
[110,123,187]). Further factors in effectiveness include the content and
clarity of videos or other tools used to facilitate reection [98], the
expertise of teacher educators [208] and the activities used to facilitate
reection [123]. It also takes time to develop skills in critical reection
and constructive peer feedback [218].
Research suggests there may be limitations to the level of reection
that teachers can reach. A study in India found teachers were able to
reect on the social, institutional and economic challenges they faced.
Yet further reection around solutions to address those problems
required more structured and scaffolded support from the TPD facili-
tator (e.g., through prompts and probing questions) [123]. The use of
more active types of technology such as digital diaries, storytelling [79,
118] or blogging [92,169,208] may afford a deeper level of critical
reection, while maintaining the social dimension of peer interaction
[154]. However, teachers levels of participation in blogging may be
largely inuenced by cultural factors, time, enforcement by school ad-
ministrators, and their need for an online community [91]. While
blogging supports teacher networking, traditional institutional cultures
may prioritise teacher independence over collaboration [92].
5.1.4. Using technology to improve subject knowledge and subject pedagogy
Software applications to improve teachers pedagogical content
knowledge [200] were predominantly used in mathematics (f=19) or
science subjects (f=15) (e.g., [30,97]). This compares to a total fre-
quency of 35 for all six other subjects mentioned. A series of studies of
pre-service teachers in Turkey indicated the particular promise of reg-
ular use of open-source geometry software ‘GeoGebra, used in 190
countries to date. In conjunction with certain pedagogies and support
structures, use appears to improve teachers conceptual understanding
across a range of topics [9,99,213,256,257], although controlled studies
are rare. Similarly, using other software has been found to improve
teachers knowledge of geometry [203,231].
10
Nevertheless, evidence
suggests that despite the potential for software to support teacher
learning, especially in science and mathematics, its scope may be limited
to specic concepts or theories, without evidence of changing practice.
Also, initial weaknesses in subject content knowledge may preclude
more than supercial uses, again inhibiting impact on teacher learning
[97].
TPD models are more likely to effectively facilitate changes in both
subject knowledge and pedagogy when teachers have opportunities for
creation of digital teaching and learning materials (theme extracted
from the literature reviewed). Studies specically referenced the use of
OERs(f=20) and editable web pages (blogs and wikis: f=14). TPD
models that use tech-supported problem-based learning [86] and
collaborative design teams [2,82,83] have proved effective in improving
teacherstechnological pedagogical content knowledge (TPACK) and
digital skills.
11
The (under-researched) role of the teacher educator
and the accompanying support materials are paramount for success, as
measured through improved classroom technology use, teacher tech
skills/awareness, and/or TPACK. Nevertheless, the effectiveness of
computer-mediated TPD models is often shaped by teachersprior
training in computer use, and the time they spend using computers [57].
A key factor for success in these tech-based TPD models is teachers
collaborative creation and sharing of digital lesson and TPD resources
through experiential learning (e.g., [58]). On the other hand, experi-
ential learning may be less effective in developing teachers TPACK in
the absence of rich practical teaching experiences [168].
Empowering teachers to collaborate, create and share resources has
the potential to engage teachers, foster agency and promote TPD in low-
resource contexts in particular. Installing computers in schools may
enable teachers to develop new digital skills (e.g. Nigeria, [153]) and
access web-based materials, overcoming challenges in contexts where
other teaching and learning resources are scarce (e.g. Rwanda, [151]).
Technology can provide TPD to stimulate teacherscreative use of both
digital and non-digital resources. Examples include video-based TPD in
rural Nicaragua [111] and multimedia player use in rural Malawi [24].
Research emphasises the importance of enabling ownership and
agency in the design processes ([57,111,186]) and providing teachers
with adequate time to learn how to effectively use technology for con-
tent design [49]. Furthermore, knowledge sharing on open licensing was
found to positively inuence teacher adoption and adaptation of OERs
([65,249];). The potential benets of content creation to enhance equity
include empowering actors to create contextually-relevant resources
that represent marginalised groups (see Section 5.2 for more on equity).
Nonetheless, few studies explored this and many suggested that the
localisation of content using technology can be complex,
time-consuming and resource-intensive [159,245]. Further, larger-scale
research is required to better understand the potential that technology
can offer to support teachers development and foster inclusive
practices.
Pre-loaded devices encouraged experiential learning opportunities in
17 (10%) of the 170 studies. The purposes of these models for teachers
were often twofold: to experiment with exemplary multimedia resources
provided in the device, and to create their own resources (School-to-
School International [[189], 2017 [56];]. Nonetheless, evidence was
again limited to self-reported data rather than observational changes in
classroom practice. One exception to this is the jiFUNzeni blended
learning approach described in Exemplar 3.
Exemplar 3
jiFUNzeni blended learning: Creating EdTech resources for learner-centred pedagogy
The jiFUNzeni approach utilises solar-powered tablets, OERs and open-source software to provide teachers with opportunities to create
multimedia resources (PDFs, video clips, podcasts and images). These are either embedded in HTML content or electronically published [157,
158]. This initiative enabled teachers in Kenya to build instructional capacity across subjects through the thoughtful development and delivery
of relevant content. One year after the intervention, follow-up interviews with teachers and TPD facilitators revealed teachers still used
cooperative learning and activity-based learning strategies. Key factors for success and sustainability of this programme included selecting
appropriate technology and viewing teachers as creators of innovative learning content.
10
In LMICs especially, there are various contextual factors that may shape the
effectiveness of software use (e.g. teachersICT skills, gender, prior knowledge,
access to internet connectivity, quality of the software design: [231]). However,
in neither of these articles were these contextual factors addressed.
11
In these studies, technology is used as a modality to support teacher
learning, but the role of the teacher educator or TPD facilitator is unexamined,
albeit pivotal in ensuring the technology is used effectively. Problem-based
learning, for example, requires a high level of pedagogical knowledge of the
facilitator. Tech is simply the delivery system for more difcult tasks such as
building teachers critical thinking skills.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
18
Despite teachers subject content knowledge being a popular inten-
ded outcome of the TPD models reviewed, there were notable method-
ological limitations. Only 23 (35%) studies with subject content
knowledge as one of the TPD outcomes utilised teachers test data to
assess their learning.
5.1.5. Using technology to foster exible learning environments
Technology has facilitated new, exible and innovative approaches
to TPD, particularly for pre-service teachers, including through inde-
pendent or self-study TPD (f=22), blended learning (f=32), or virtual
learning environments (VLE) (f=33). Technology access can motivate
teachers to pursue self-study TPD opportunities [149] and to remain in
these programmes until completion. Experimental studies have found
that the exible nature and enhanced interaction offered by blended
learning environments improve teachersknowledge of a range of
topics, including ICT (e.g., Yemen, [179]), science (e.g., Pakistan,
[135]), and classroom management theory (e.g., Turkey, [103]). Flip-
ped classrooms, for example, allow teachers to watch lectures or engage
with other multimedia content off-site before discussing them in
teaching sessions; increasingly popular in HICs, these proved less com-
mon in LMICs (f=7). Flipped models allow time to use more participa-
tory teaching and learning methods in lectures [135] (where lecturers
know how to use these methods). They enable lecturers to select the
combination of virtual and in-person activities that maximises student
engagement and learning [103]. Despite these benets, evidence and
explanations for changes in practices remain limited in many studies
reviewed.
VLEs (also referred to as Learning Management Systems or LMS)
the most common of exible approaches have been used increasingly
in ipped, blended or distance TPD models. Like social media, carefully
designed VLEs can provide rich learning opportunities outside of the
physical classroom. VLEs can connect pre-service teachers with the
lecturer outside of class and make it easier to follow teaching content
and assignments or access additional study materials and learning ac-
tivities [13,52]. While VLEs can be used for direct transmission, they can
also be used to foster collaborative e-learning environments. Such en-
vironments can support the development of teachersproblem-solving
skills and metacognitive strategies for self-regulated learning, and in-
crease motivation [146]. Nonetheless, individual factors such as teach-
ers epistemic beliefs (on how knowledge is developed) and attitudes
toward tech may mitigate VLEs effectiveness (ibid). More experienced
or motivated peers or even adolescent students who may have more
advanced ICT skills than their teachers [159] may serve as an impor-
tant support. This can increase teacher engagement, facilitate teacher
learning with ICT [147], and ultimately foster changes in classroom
practice [238]. In the latter study in a Chinese secondary school, an
online community supported structured activities focused on collective
lesson planning, individual reection, peer critique and resource sharing
by teachers of the same subject.
5.2. Using technology-mediated TPD for more equitable systems (RQ3)
A global focus on educational quality and enrolment rates has shone
a light on the magnitude of disparities existing within and between
education systems. This has led to recognition that teachers are a
foundation of an inclusive education systemand require support to
develop skills to cater for needs of all learners ([227], p. 153). This
section explores how technology can help foster more equitable edu-
cation systems (RQ3, Table 1) a common goal of education reforms,
but practice often lags within TPD initiatives, as we demonstrate below.
5.2.1. Using tech to provide TPD through remote learning
Technology has the potential to widen access to professional learning
opportunities for teachers remotely. The literature relating to remote
learning contained research studies where distance (f=30), blended
(f=34), remote/rural (f=23), and Massive Open Online Courses
(MOOCs) (f=10) were referenced and coded. The literature generally
included little detail on demographic or socio-economic characteristics
of teacher participants, and rarely disaggregated data regarding inter-
vention effects. Moreover, why teachers engaged in remote learning and
how adaptation could meet diverse needs remained unevidenced.
Remote learning can provide access to TPD sustainably throughout
teachers careers, for instance through MOOCS [89,167,237,248].
Different instructional design frameworks are required for different
learner types, though. For example, more experienced teachers may be
more open to experimentation [237]. Linguistically- and
culturally-adapted MOOCs have been found to contribute to high
participation and completion rates [246]. Many initiatives highlight the
importance of ensuring local support and engagement, which are central
factors in effective MOOCs based in India (ibid) and Uganda [167].
A prominent challenge for purely remote TPD is the lack of oppor-
tunity for teachers to apply learning, agged earlier in Kok & Blignauts
[98] study of 15 teachers in rural communities in Namibia. Participants
there reported that the DVD-based TPD afforded no chance to apply
theoretical knowledge on teaching learners with Special Educational
Needs and Disabilities (SEND). By contrast, Henaku & Pobbi [70] found
minimal differences between the classroom management skills of Gha-
naian teachers who undertook distance learning programmes versus
traditional in-person TPD. The large sample (n=500) analysis was based
on self-reported surveys rather than observational data, but ndings
were triangulated via head teacher assessment, mitigating the method-
ological limitations somewhat. There are also tensions around the
structure of distance-based TPD. For instance, Sri Lankan and American
teachers in the international collaboration initiative previously dis-
cussed found that highly regulated and frequent assessment could make
interactions too formal and reduce teachersfreedom of expression
[190]. Conversely, Harley & Barasa [65] observed that the highly
structured version of the Teacher Education in Sub-Saharan Africa
(TESSA) programme was effective in distance education programmes
(see Exemplar 6). Finally, logistical issues for remote learning can
impact teacher motivation and learning, evidenced through Mokoenas
(2017) research on 65 pre-service teachers in South Africa [140]a.
Communication problems with the teacher education ofce and a lack of
placement visits demotivated teachers.
Blended learning programmes can provide teachers with opportu-
nities not possible through purely distance or in-person TPD. The
Teacher Education through School-based Support in India (TESS-India)
programme used a combination of digital forums and physical learning
spaces to merge global knowledge and local support; the MOOC
completion rate increased and classroom teaching practices improved as
a result [246,248]. The opportunity to apply new knowledge and
pedagogical practices in the classroom and later reect (virtually/in
person) with peers and facilitators resulted in improved teaching prac-
tices in Bangladesh [196], Kenya [159] and Pakistan [7].
However, issues with blended TPD initiatives included pre-service
teachers in Turkey reporting that certain aspects of the online environ-
ment made them feel isolated and hindered learning Yılmaz and Malone
[254]. Teachers in Indonesia had low levels of participation in online
learning, due to various factors including reduced commitment to the
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
19
programme because of the absence of formal consequences.([243], p.
388). Again, motivational and social dimensions of technology use need
to be carefully considered to ensure blended TPD is effective (see Section
5.1.5). Moreover, experimental designs examining the relative impact
across in-person, blended and remote TPD models were uncommon.
Exceptions include previously discussed studies by Kotze et al., [100]
and Qasem & Viswanathappa [179].
Providing TPD in rural settings has potential to signicantly address
equity issues; nonetheless, few studies were coded both remote/rural
areaand either distance(f=3) or blended(f=4). Generally, only 23
papers (14% of total) involving teachers in rural/remotecontexts were
identied. Thematic coding of the research highlighted that rural con-
texts can often be associated with complex issues and other forms of
marginalisation. Moreover, while some research designs included both
urban and rural schools (e.g., [159,175,237]), few detailed dis-
aggregated effects and distinct causal pathways of impact. The impor-
tance of separating out effects cannot be emphasised enough. Ongukos
[159] study in Kenya involving a needs assessment, for instance,
discovered that urban teachers preferred training on assessment,
whereas rural teachers wanted TPD for teaching large classes. Tailoring
the TPD content across contexts subsequently contributed to
longer-term pedagogical improvements.
Studies also detailed TPD initiatives utilising innovative technology
solutions to infrastructural challenges faced in rural/remote contexts.
This included enabling access to TPD materials and facilitating com-
munities of learning through social messaging applications (see Section
5.1.2), solar-powered devices to combat electricity shortages [159], and
using a local Wi-Fi network connecting teachers to a central server [72].
Researchers emphasised the importance of providing support to teachers
in rural contexts, for example, in troubleshooting technology issues,
enabling reection and providing motivation [100,111,160,248].
The exibility provided through remote learning via mobile devices
with pre-loaded content was another advantage agged in multiple
studies in rural settings (e.g., [24,197]). Many studies, however, were
small-scale and/or based on pilot initiatives. Others lacked detail around
the feasible limits for device sharing, rendering it difcult to draw
conclusions for sustainability or scalability. TPD programmes that
require only one device or a handful per school are commensurate with
common resource constraints. However, teachersbuy-in to minimal
technology sharing among peers can be limited, precluding experi-
mentation afforded by individual devices [198]. Research also demon-
strated how technology can sometimes be ineffective in providing TPD in
rural settings [28,98]. Rana et al., [181], for instance, detailed how
bringing technology into rural Nepali primary schools already experi-
encing infrastructural difculties (from earthquake damage) created
further organisational complications that compromised TPD effective-
ness. These studies highlight the importance of ensuring constructive
and contextualised application of technology to support teachers in rural
settings.
5.2.2. Using tech to improve language of instruction
Teachersknowledge and skills in their language of instruction can
impact teaching quality, and technology can be used to improve this.
EAL/SLL was a focus in 45 studies reviewed. While studies noted lan-
guage benets of technology use, few measured actual improvements in
teachers language development or teaching practices relating to lan-
guages. There was also no literature relating to the use of technology to
assist TPD for sign language, which is concerning, considering 34
million children globally have a hearing disability [227].
A number of studies detailed how technology enabled greater lan-
guage practice opportunities, including delivering TPD in a target lan-
guage [81,190,198]. Participants across studies were found to have
increased their language condence and competency through wider
language exposure. This included providing second/additional language
TPD content on mobile devices as in the English in Action initiative in
Bangladesh [198]. Further examples were participation in social media
groups (Bett & Makewa, 2020), using language applications [20] and
reading and writing practice via the creation of e-portfolios by
pre-service teachers in Malaysia [81]. E-portfolio applications online
platforms where teachers can access elements such as lesson plans and
feedback, peer forums, and exemplar lesson videos were also found to
meet the planning needs of pre-service English teachers in Turkey, but
they were perceived as less effective in supporting teaching and reec-
tion [88]. Further, research rarely documents teachersinitial language
levels or any disaggregated initiative effects; these measures could
inform adaptation or tailoring of language content.
Using video reection to self-assess language competency was a
feature of three studies [59,123,218]. All noted that teachers used the
videos of themselves to identify language errors, such as grammar and
pronunciation mistakes, hence improving their language skills. How-
ever, none of the studies explored the impact of sustained video reec-
tion to improve language competency over time. Virtual coaching (see
Section 5.1.1) for language learning offers further potential benets: in
Kotze et al.s [100] large-scale study cited earlier, it presented signi-
cant cost advantages, which is particularly important when providing
TPD in multiple languages. Nonetheless, there were indications of po-
tential negative effects from neglecting home language skills after a
year, which raises concerns around the trade-off to improve learners
English competency. Furthermore, the positive impact on teachers
English prociency was an additional, unexplained benet of the
intervention.
5.2.3. Using tech in TPD to improve learning for the most marginalised
Technology presents both opportunities and challenges for support-
ing teachers to adapt practices for diverse learnersneeds. Research
highlights the importance of working with teachers to understand their
priorities and nuances in challenging settings to more effectively tailor
support (Anamuah-Mensah et al., 2013 [159,160];). Nonetheless, this
review found a signicant lack of technology-supported TPD initiatives
focusing on the specic needs of either marginalised teachers or learners
(see Section 4.2.4.2).
5.2.3.1. Marginalised teachers. Despite the potential reach of technol-
ogy designed to be inclusive, there was little evidence on tech-based TPD
for SEND teachers. Wormnaes & Sellaegs [253] study in Uganda found
that the audio-described educational video material was valued by the
12 visually impaired participating teachers. It led to constructive
reection as teachers considered their involvement and emotional
engagement with SEND learners. The research also discovered that
when paired with sighted teachers, visually impaired teachers partici-
pated less in discussions. It is critical that such dynamics are considered
and effectively addressed to ensure TPD is equitable.
Only three studies explored the role of technology in supporting TPD
for teachers in refugee/displacement contexts, where language barriers
and emotional trauma commonly pose specic needs. Mendenhall
et al.s [133] multifaceted approach is one example (Exemplar 4).
Language applications have enabled female Syrian refugee teachers in
Lebanon and Sweden to undertake self-directed learning and simulta-
neously improve new language skills and teaching practices [20]. There
was also evidence demonstrating the benets of co-designing TPD ini-
tiatives with marginalised teachers. For instance, Kennedy & Laurillard
[89] worked with Syrian and Lebanese teachers to construct a MOOC
designed for TPD in education in emergency (EiE) contexts. The logic
that developing EdTech initiatives in the most challenging contexts and
extending them to more favourable conditions is compelling but rarely
implemented.
12
12
See https://blogs.worldbank.org/edutech/scaling-up
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
20
5.2.3.2. Marginalised learners. Despite the limited literature identied
concerning learners with low socio-economic status (11 studies, 7 of
which involved rural/remote learners too), evidence is emerging that
tech-mediated TPD can be effective in such settings. The use of partic-
ipatory video-making in South Africa, for example, was a valuable tool
in enabling teacher agency and opportunity for group reection on
matters of marginalisation, such as poverty, orphans and HIV/AIDS
[137]. Teachers using laptops, video and OER in the OER4Schools
initiative in Zambia were found to have raised their expectations of what
rural and vulnerable students can achieve [72,73]. They developed their
awareness of all learnersprogress. The initiatives holistic approach
renders it difcult to separate out the effects of specic technology.
However, the programme highlights the importance of a multimodal
and blended approach to effectively support and empower teachers to
foster inclusion.
There was a near absence of any TPD initiatives focused on girls or
learners from SEND or EiE groups. One ineffective DVD-based remote-
learning model for teachers of SEND learners in Namibia [98] is dis-
cussed previously. In a rare study including both pre- and in-service
teachers, MacEntee [118] explored the contributions and challenges
for participatory visual methodologies to enhance HIV education in
rural schools across South Africa. Results indicated that the
photography-based methods covered gender themes and facilitated
teachersreexive learning and the production of local resources in
under-resourced schools. Challenges, however, included limited tech-
nology access (the research project supplied all digital tools) and po-
tential to evoke trauma and discomfort among teacher participants.
Another study, at the intersection of gender and EiE, sought to demon-
strate the peacebuilding potential of positive gender socialisation in a
conict-affected region of Uganda [28]. This study found no clear evi-
dence of positive impact from SMS messaging on teachers attitudes,
knowledge or practice (in contrast with studies in other contexts, re-
ported in Section 5.1.2).
The ndings in this section emphasise the importance of TPD to cater
for diverse learning needs, yet among the identied literature there was
little detail on how TPD targeted this. The evidence suggests that tech-
nology has the potential to facilitate TPD designed to support margin-
alised learner needs; however, the complexity of these needs requires a
holistic approach.
5.3. Multi-level factors inuencing tech-supported TPD (RQ4)
This section explores the intersecting, multilevel factors inuencing
tech-supported TPD initiatives within a coherent education system. It
builds on Figure 1 and shows how different elements are informed by
government, institutional and teacher priorities. In doing so, it answers
RQ4: in what ways are TPD initiatives systemic in their approach,
considering macro-, meso- and micro-level factors during design and
implementation?
5.3.1. Macro-level factors inuencing tech-supported TPD
Macro factors manifest at the national and sub-national level.
Despite the importance of this level for governing and coordinating
effective TPD systems, 25% of studies did not discuss system-level fac-
tors. This subsection discusses the main themes emerging across the
other 75% of studies.
5.3.1.1. Resourcing TPD in a fractured environment. Infrastructure is a
cross-cutting element in the TPD system. At the macro level, infra-
structure involves the construction and resourcing of schools with
qualied teachers, equipment (including tech devices), teaching and
learning materials and more, by government. It also includes the
availability of reliable power and connectivity, crucial to successful
implementation of most tech-supported TPD interventions. Macro-level
infrastructure was the most frequently named code among system-level
factors, appearing in 24% of studies (f=41).
In a study investigating the integration of tech for TPD in Nepal, lack
of government funding and investment in digital infrastructure meant
that schools could not implement educational directives that govern-
ment policies set out. These included the provision of ICT infrastructure
and professional development of university staff to integrate ICT in
teacher education [182]. This lack of funding, and so government
ownership over the Nepali TPD system, created a gap in service delivery,
lled by international organisations. This example demonstrates a
fractured picture of TPD implementation that is all too common [228].
Onguko [159] corroborates, highlighting the importance of coherence
to ensure a degree of complementarity between implementing parties,
funders, governments and those most affected (i.e., teachers and stu-
dents). To achieve coherence around national priorities, governments
must coordinate and manage partners closely. A push to promote a
learning culture across the TPD system would avoid unnecessary
duplication and help large-scale initiatives to learn from both mistakes
and successes of small-scale pilot projects([171], p. 11). This serves to
join up a disparate system of initiatives, actors and funding for effective,
equitable and sustainable TPD to promote improved learning outcomes.
5.3.1.2. National policies, assessment frameworks, accountability and
incentives. The policy or political environment was the third most
frequent code among factors at the system level(37 or 22% of
studies).
13
The role of policy in stimulating widespread change across
the TPD system is exemplied through Chinas educational informa-
tionizationgoals. These aimed to connect every school to the internet
and integrate ICT into the curriculum by 2010 [109]. Projects set up to
achieve these goals included a distinct focus on developing studentsand
teacherstech skills. Increased policy attention on EdTech use and
tech-supported TPD effectively addressed both teachersand learners
needs. The TPD improved instructional delivery and more creative uses
Exemplar 4
Mobile mentoring in a refugee context [133]
Teachers in the Kakuma refugee camp in Kenya took part in a TPD initiative and were provided with mobile phones, data and airtime. Over one
year, they took part in workshops, virtual peer coaching and mobile mentoring. The peer mentoring fostered communities of learning and
relationship building in the camp, and the global mentoring enabled two-way knowledge sharing as external mentors were able to understand
the context and tailor support. The mentoring provided real-time responses to challenges and enabled feedback loops (both between mentors/
mentees as well as project staff). Teachers also reported using the mobile devices to (a) access the internet and engage in self-directed TPD, for
example to nd content on how to support SEND learners, and (b) share innovative teaching practices through photos or exemplar videos.
Further research might investigate the impact on their students.
13
Although there are myriad factors at the policy level that shape the effec-
tiveness of tech for TPD models (e.g. teaching standards and qualication
frameworks, recruiting and promotion mechanisms, pay, compensation and
tenure), this section focuses on those factors most common in the literature
reviewed.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
21
of lesson materials, in turn promoting improved learning outcomes.
However, teacherspedagogical beliefs determined how the technology
was used, for example whether teachers creatively added animations
and images to make their lessons more vivid(ibid., p. 114). Moreover,
teachers were reluctant to use technology when faced with mounting
exam pressures, reverting to traditional, non-tech practice (ibid.). A
decade on, the policies and projects that made up Chinas educational
informationizationled to considerable national achievements, such as:
development of substantial digital resources, including MOOCs.
Large-scale TPD on
EdTech has also reached tens of thousands of teachers [239].
National assessment frameworks are central sources of account-
ability; indeed, teaching to the exam is arguably an inevitable
consequence of high-stakes assessment [109]. Exams need to reect
broader technological shifts for teachers to value and normalise inno-
vative classroom practices using technology (ibid.). Macro-level
assessment frameworks and curriculum uptake have clear impacts on
in-classroom teaching and learning, particularly in a pressured envi-
ronment. Kotze et al., [100] found that inadequate curriculum imple-
mentation in South Africa triggered teachers avoidance of more
complex (often learner-centred) activities. Comprehensive and sustain-
able TPD systems are a key method of breaking the cycle and ensuring
policy change stimulates classroom-level change [181]. In sum,
high-quality TPD is a crucial link between policy and practice, poten-
tially enabling teachers to maximise the capacity of technology to
improve classroom pedagogy [171].
The essential underpinning to a coherent TPD (and wider education)
system, is the effective collection, use and interpretation of data. Timely
and accurate data allows decision-makers to make informed decisions
[11]. Data span levels and topics from individual teacher attendance or
student assessment data to school-level data on performance or infra-
structure, through to national school system data. However, data are
generally managed, used and interpreted at the macro level.
Data are used less frequently at the micro- or meso-levels to support
teachers or institutional leaders. An exception is the work of McKenney
and Mor [130], who provide design guidelines to create a valid, prac-
tical software tool that could support teachers in collecting, annotating,
interpreting and analysing data in their classrooms. These include
encouraging teachers to test their own conjectures about effects of their
learning material designs on student progress, and to reect with peers
on the data and its implications for their conjectures.
Exemplar 5 also exemplies how data can be used to enhance the
coherence of a TPD system, connecting different actors: reciprocal
accountability affords micro-level actors the space to make their voices
heard.
Finally, macro-level working conditions can impact teacher moti-
vation [111]. In particular, consistent salary payments and career pro-
gression opportunities are basic elements of a teachers social contract
that affect attitudes towards the profession and engagement in TPD.
Awareness of these contextual realities, including incentives (e.g.,
[177]), time constraints and teachersworking conditions, can help
ascertain the reasons why teachers do (or do not) engage in TPD.
5.3.2. Meso-level factors inuencing TPD
The meso level involves schools/colleges and communities, coaches
and facilitators, TPD initiatives and tech support, and more. It plays a
vital role in linking the micro and macro levels. This section discusses
meso-level constraints, collective culture and teacher agency, and the
role of school leadership (most research focused on schools rather than
pre-service education settings; however, messages may apply there too).
5.3.2.1. Meso-level infrastructure and institutional constraints. Meso-level
infrastructure can comprise school equipment, a schools technological
resources, localised power/energy supply (e.g., solar mini-grids, [255]).
This infrastructure is often provided by the state or other macro-level
actors, however equipment, management, maintenance and use typi-
cally sit with schools and individuals.
Dlamini & Mbatha [40] note that uptake and usage of the extensive
school-level ICT infrastructure in South Africa is low. This is due to a
lack of school-based TPD equipping teachers with the necessary skills to
use technology effectively. As such, the authors advocate for TPD op-
portunities supporting teachersuse of tech to foster the skills, knowl-
edge and enthusiasm to integrate technology into everyday teaching
practice. Hence, coherence between provision of technology to schools
and TPD concerning how to use these technologies in schools is vital
[181,182].
Institutional constraints can mean meso-level physical constraints,
such as a schools lack of access to technological hardware and/or de-
vices, or large class sizes affecting studentsinstructional time. Institu-
tional constraints can also denote more ‘humanfactors, for example,
triggering an unenthusiastic school culture ([1], p. 102). Such con-
straints can limit the creativity of teachers, even those enthusiastic about
using technology. Where teachers must share limited technological de-
vices, for example, between colleagues or students, the potential added
value of the technology has a ceiling as even enthusiastic teachers have
limited access. These constraints, whether ‘physicalor ‘humancan
create perpetuating cycles of despondence concerning engagement in
TPD activities, ultimately negatively impacting learnerseducational
experiences.
5.3.2.2. Collective culture and teacher agency. Teacherscollective cul-
ture can take the form of a national teaching workforce (macro), or the
embedded culture within a given school or cluster of schools (meso).
Collective culture is a further signicant factor in determining teachers
attitudes towards, and participation in, TPD programmes and their
overall perceptions towards teaching as a whole. Yet a mere 12% (f=20)
of studies discussed cultural dynamics.
Lindenberg et al. [111] found that teachers who distrusted the
Nicaraguan education system due to previous state failings were less
Exemplar 5
Tusomes National Tablets Programme in Kenya
A dashboard compiling data on the numbers of observations (recorded on tablets) that coaches undertook was developed to enhance the
accountability structures within the ministry of education. This was targeted at establishing greater links between the sub-national (county) and
national levels [172]. In addition to heightened accountability structures between coaches, country programme directors and government,
coaches increased use of tablets improved the quality of their instructional support to teachers. However, ‘modest relationshipsbetween
increased coach visits and overall improved learning outcomes demonstrated the complexity of making system adjustments to improve learning
outcomes. The research ndings were used to adapt the TPD offering. It highlights the need for further research exploring the relationship
between TPD, accountability and learning.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
22
motivated to engage in the TPD. That said, those teachers who did
engage in the TPD demonstrated a degree of change in their teaching
practices. Yet, even the most effective, equitable and sustainable
evidence-based, tech-supported TPD offerings with all the ‘essential
ingredients that would generally denote success may not work in a
given context if teachersattitudes are negative towards their education
system as a whole, or even towards their mentors, coaches or teacher
educators. Mutual trust and condence are essential for success.
School culture dened as the basic assumptions, norms and
values, and cultural artifacts that are shared by school members([126],
pp. 89) has the potential to both enable and constrain teachers [1].
School cultures can afford teachers with the requisite experience, safe
space and condence to be creative in the design, delivery and assess-
ment of learning. Yet, school cultures can also often pose signicant
constraints on a teachers creativity, inhibiting their inherent urges to
go off scriptin order to really engage students [45]. As such, providing
teachers with new analytic and design tools to overcome institutional
constraints by taking on diverse and changing roles during TPD activ-
ities can shed light on the embedded institutional (often ‘naturalised, or
deep-rooted) behaviours that may constrain teachers in the pedagogical
process ([12], p. 41). Technology can play a signicant role in drawing
in remote, external peersperspectives. It removes teachers from their
school echo chambers, injecting the perspective of the ‘other to learn
and grow from and to avoid reinforcing only local traditions.
5.3.2.3. School leadership. School leaders play a large role in ‘culture
settingat schools and in wider communities; yet just 20 (12%) studies
discussed the role of leadership support in TPD. A school leader who can
properly manage and maintain tech resources and provide teachers with
the adequate support to integrate technology into classroom practice
can act as a lynchpin for the school [22]. Teachers who are hesitant
about or resistant to change need role models to show them how they
can progress their practice [40]. Agyei & Voogt [1] go further, stating
that enabling and supporting school leaders to provide proper peda-
gogical leadership in ICT integration [...] will inspire new teachers to
push the boundaries of using ICT-enhanced activity-based learning
innovation(p. 103).
Dlamini & Mbatha [40] assert that inspiring school leaders can
positively impact learning outcomes. In their study on South African
teacherstech readiness, the high-performing schools had involved
school leaders working on school ICT strategies alongside champion
teachers on the groundwho were condent modelling their practice to
others ([40], p. 28). School systems must establish mechanisms for
identifying high-performing teachers to act as these champions; 19% of
the studies cited the benets of peer support.
5.3.3. Micro-level factors inuencing TPD
Whether in the classroom, or through distance means, education
decision-makers must prioritise the micro level where learning hap-
pens when planning, designing and implementing TPD activities.
5.3.3.1. Co-design, experimentation, and teacher voice and agency. Kalo-
giannakis [84] argues that most ICT training lacks attention to the
context in which teachers work(p. 14). An effective method of ensuring
decision-makers are mindful of micro-level factors is through co-design,
where teachers are decision-makers, themselves involved in the plan-
ning and design process.
Customisation by, with and for teachers can be supported through
design-based research (DBR)
14
that is contextually responsive to
participant teachers needs [131] and sustains teacher learning [103].
An online TPD programme for STEM (science, technology and
engineering) subjects in Pakistan in participants schools allowed the
researcher and teachers to collaboratively evolve the programme
implementation and evaluation ([7], also discussed in Section 5.2.1.)
The resulting accessible online learning environment provided teachers
with opportunities to explore STEM integration by building a com-
munity of practice that integrated both asynchronous and synchronous
technology-mediated environments(ibid., p. 204). Such DBR models
can also have positive multiplier effects on teachers acting as role
models and creating positive collective cultures, elements which can
traverse into the meso level and increase potential for scalability.
Teachers are a heterogeneous group, thus the concept of affording
teachers agency must be tailored to teachersbackgrounds, experience,
beliefs, knowledge, attitudes and motivation (appearing in 99, or 58%
of, studies). One-size-ts-all TPD designs can be problematic. As Celestin
& Yunfeis [26] ndings reveal, teacherspre-training characteristics (e.
g., learning readiness and personal capacity) are signicant predictors of
post-training performance. Therefore, initiatives must embody the
common characteristics of effective TPD (as per Section 2.1) and treat
teachers as individuals (just as learners should be treated in the class-
room). De Clercq & Shalem (2014) propose a differentiated form of TPD,
whereby professional development differs in its organisation, pace,
location and modality, according to teachersneeds.
Several studies demonstrate the complexity of teacher agency versus
appropriate scaffolding for teachers and show why these features must
be treated as uid concepts. In Bangladesh, pre-loaded iPods increased
teachersclassroom use of audio and video materials, but teachers rarely
created their own resources [198]. A reliance on ready-made resources
such as scripted lesson plans may therefore inhibit teachers creativity,
potentially damaging long-term teacher motivation in the process
(ibid.).
(Semi-)scripted lesson plans can set common expectations around
student learning and provide less experienced teachers with an appro-
priate amount of scaffolding, yet more experienced or condent teachers
might nd them demotivating; responsively fading support is important.
While Piper, Zuilkowski et al., [174] found that (semi-)structured
pedagogy offered a cost-effective approach to TPD in Kenya, another
study exploring teachersguides across 13 countries and 19 projects
offered a salutary lesson. Overly scripted guides produced poorer
teacher learning outcomes than simplied ones [173]. Teachers did not
adhere to the scripts, often reducing group work for more
teacher-oriented activities. Moreover, Kotze et al., [100] suggested,
without teacher learning and teacher agency, these elements of the
instructional infrastructure [structured pedagogy, scripted lesson plans
on tablets, teacher guides etc.] have little chance of transforming the
everyday learning activities and tasks in the classroom(pp. 56).
Exploiting technology to share lesson videos or offer opportunities for
pedagogy-related discussions (e.g. on social media) can be fruitful
though (see 5.1.2).
5.3.3.3. Motivation, time constraints, teacher needs, skills and attitudes
towards tech. Motivation can manifest across the system and implicate a
range of educational stakeholders. As discussed in Section 5.3.2.2, Lin-
denberg et al., [111] speak to the high turnover of teachers in Nicaragua,
where passion for education is often second place to a teachers need for
a steady salary to support their family. As such, the positioning of the
teaching profession in certain LMICs can mean that there are more
‘push than ‘pullfactors. Continual teacher turnover has implications
for sustainability and scalability of TPD programmes, reliant on
continuous feedback loops and the transfer of knowledge, experience
and learning, enabled through retention and progression of teachers.
The mode of TPD can also impact a teachers willingness to engage in
professional learning. While blended or distance-based models may be
seen as exible forms of participation, micro-level factors such as
teachersdigital literacy or lack of access to personal devices can
obstruct participation. This was observed in the previously discussed
14
DBR involves iterative cycles of design, evaluation and re-design, in
collaboration with local stakeholders. DBR studies build theory and generate
design principles as well as impacting on practice.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
23
Widodo et al., [243] study (Section 5.2.1) on blended TPD in Indonesia
which saw very low online participation compared to the in-person
approach. Rapidly increased mobile phone permeation in LMICs may
offer more accessible opportunities, yet these are vastly under-used to
date (see Figure 9). Unsurprisingly, another key (dis)incentive of
teachersengagement in TPD is time constraints, appearing in 19%
(f=32) of studies. An online collaborative learning initiative proved
ineffective for science teachers in Botswana, with work pressures stated
as a core concern [19]. Schools can play a key role in alleviating work
pressures by supporting and creating a culture of ongoing teacher
learning in the workplace. However, schools are of course part of the
broader system that may heighten the pressures placed on schools, in
turn pressurising teachers.
It is uncommon for technology-mediated TPD programme designers
to assess the learning needs of teachers [26]; although 28% of the studies
discussed teachersprior ICT skills, this gure seems low given the
signicance of determining teachersexisting skills before implementing
a TPD initiative. Nevertheless, one study in rural Kenya found that
creating locally-relevant content following a needs assessment ensured
teaching practices were developed with support from local TPD experts
([158], see Exemplar 3). This was implemented in a blended learning
approach based on reective practice. Moreover, if teachers have
developed skills and positive perceptions in relation to technology,
students themselves may feel more condent to engage in the subject
content [26]. Conversely, negative perceptions can lead to reluctance.
Therefore, tech readiness and adoption sit rmly within a teachers
willingness, skill level and pedagogical belief system [43], along with
TPACK [66] proving crucial to the degree of creative classroom tech-
nology use [1].
5.3.4. Cross-cutting elements in a coherent system: cost effectiveness,
sustainability and scalability
All elements of the TPD system are co-dependent. This section
highlights tech-supported TPD studies through this multilevel lens.
These designs are generally cost-effective, scalable and/or sustainable
(Exemplar 6 provides the rst example).
5.3.4.1. Cost-effectiveness. The cost-effectiveness of a TPD initiative and
its related potential to scale is integral to building and expanding TPD
initiatives. It is particularly important to understand the costs involved
given the scarcity of resources that constrains TPD in LMICs [24].
However, cost analysis is complex, particularly when considering
broader costs beyond monetary value. Though this complexity must be
recognised, only 11% of studies reported any detail of cost. This reects
the wider issue discussed by Bruns et al., [22] of a huge publication
bias; programs with negative or negligible impacts are almost never
reported, and among programs with positive impacts, cost data are not
always reported(p. 226).
Handheld devices can be more versatile than personal desktop
computers and are favoured by teachers, as seen in the Primary Math-
ematics and Reading (PRIMR) Initiative [171]. Not only do portable
devices enable more exible access, but they are also less costly per unit
and provide greater opportunities to be shared among colleagues. This
positive move away from the locked, gatekeeper-controlled [computer]
lab(ibid., p. 11) denotes a shift towards technological integration
within classroom settings. However, data from Figure 9 show that the
TPD landscape does not tend to follow this shift, with 42% of all men-
tions of devices focusing on computers (although laptops are included
within this category). That said, regional data analysis as illustrated in
Appendix F (Figure 12) shows that mentions of smart devices and
computers for TPD are equal in SSA and South Asia. A further analytical
lens (time) shows that computers made up a slightly higher share of
devices in studies published between 20082013 (48%) compared to
studies published between 20142020 (40%).
The PRIMR study elucidated cost implications across different levels
of the education system. For example, providing tablets at the meso level
(to tutors) as opposed to the micro level (to teachers or students) keeps
costs manageable. In terms of learning outcomes, all (tech and non-tech)
treatment groups performed better than control groups [171]. However,
it was unclear whether PRIMRs instructional approach was the more
signicant factor versus the presence of technology.
Technology cost analyses must be transparent, comparable and
considered alongside the learning benets of any given technology [24].
Use of low-cost MP3 players for teacher learning in rural settings has
improved classroom instruction (ibid). The shared use of media players
within a school proved the most cost-effective option. Although no
learning outcome data were provided, per-lesson costs would be even
less at scale than printing paper-based materials (ibid).
5.3.4.2. Scalability and sustainability. ‘Scalabilitydenotes the potential
of a tech-supported TPD intervention to expand and be implemented
beyond the initial location(s), reaching more teachers and learners.
15
‘Sustainabilityrefers to the durability of a TPD initiative beyond its
initial time frame. Both are core to establishing coherence and
improving learning outcomes.
If a TPD initiative were to reach national scale and sustain this it
would support the development of a coherent TPD system with the
leverage to engage myriad teachers at once across different contexts,
ensuring teachers are connected in a lasting way. The national-scale
PRIMR study in Kenya (discussed in Section 5.3.4.1) is one example of
a country-wide operation which can build on the learning from previous
years [172]. However, cautionary lessons concerning simplistic as-
sumptions about scaling are learned from studies like that of Kraft et al.,
Exemplar 6
TESSA
The TESSA OER project in sub-Saharan Africa has merged educational theory, digital technologies and teaching practice [65]. The project
reached 300,000 teachers, holding promise for scaling TPD across SSA. The critical indicators of success include project ‘take-up in diverse
settings and the signicant impact on the practices and identities of teacher educators and teacher-learners. The programme altered teachers
perceptions of their role from using ‘chalk-and-talkmethods to becoming creative facilitators of learning’’ [247].
Studentslearning experiences grew through the programme. There was also some degree of impact on ministries of education, through policy
reforms and changes in practice. For example, faculties began to work more closely with schools, and materials were used to upskill school
leaders. Notably, impact was achieved despite limited ICT infrastructure and expertise.
15
‘Scalabilityusually refers to at least district-level expansion, although there
is no agreed denition.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
24
[102] on teacher coaching. This showed that effects from effectiveness
trials of larger programmes can be far smaller than the effects of smaller
programmes.
Claims that programmes are sustainable are not often validated
internally or externally [82,90]. Furthermore, a minority (18%) of
studies were longitudinal (i.e., containing more than two data-collection
time points). Moreover, follow-up after programme completion is rare;
evaluations usually take place immediately post-TPD, with little evi-
dence on the long-term effects and changes in teaching practice in these
instances. Factors such as implementation delays or unstable political
contexts (e.g., [28]; School-to-School International [[189], 2017] may
contribute to sustainability issues in LMICs.
Studies
16
that have considered sustainability in-depth include Cil-
liers et al., [29], who found that the virtual coaching mode initially
appearing successful [100] became less cost-effective than in-person
coaching after 3 years; the on-site mode was costlier but had signi-
cantly greater impacts on learning outcomes. Moreover, home language
literacy of the virtual coaching group was again crowded out and
actually declined. Additionally, a year-long OER4Schools programme
(Exemplar 2) in one Zambian school was followed up through teacher
interviews after 18 months [69]. The programme became
self-sustaining; previous participants became peer facilitators and
teachers had further developed their interactive teaching strategies. Key
sustainability mechanisms included culturally sensitive and participa-
tory programme development, semi-structured multimedia materials,
and supportive institutional and national structures for professional
learning.
Exemplar 7 presents a nal example, revisiting the HALI programme
highlighted in Exemplar 1. An earlier study undertook detailed cost
analysis and was designed to be both scalable and sustainable. These
core considerations were noted as proving key to thinking systematically
about TPD and engaging with actors, elements and processes at the
micro, meso and macro levels.
6. Conclusions
The increase in numbers of studies published over the past decade,
compared to the few from the preceding years, indicates that pioneering
research is currently taking place in this exciting eld. This review is
thus timely. It contributes to the eld by offering a wealth of examples
drawn from an unusually large corpus of 170 studies that evidence and
illustrate how we can exploit the signicant potential that EdTech offers
for TPD in LMICs. This includes how EdTech interacts with contextual
inuences and how initiative outcomes can be effectively evaluated and
better understood. It is accompanied by a truly unique resource for the
eld: an open, substantial database including ne-grained thematic
coding and quality assessment, which can support many further research
inquiries.
Seven key conclusions emerge from the ndings, as follows. These
relate to our rst three RQs; conclusions regarding RQ4 are embodied in
the Recommendations below.
1 TPD for technology use is generally under-researched in LMICs
(RQ1 characteristics of studies). Most countries produced no
research studies at all over the 12-year period. Indeed, only 5
countries had 10 or more studies China, India, Kenya, South Africa
and Turkey. These were all MICs, reecting the inequitable distri-
bution of research. This is a signicant inuencing factor, given that
what is effective in MICs may not be effective in LICs, particularly for
marginalised groups whose needs are largely ignored in the literature
(RQ3). More support for research conducted in LICs is called for.
2 Methodological issues abound: sustainability, sampling, access
to evidence (RQ1). The absence of follow-up studies makes it very
difcult to judge sustainability, and future research must consider
the long-term impacts of TPD interventions leveraging digital tech-
nologies. The majority of studies included in this review involved
small sample sizes across a limited geographical area (e.g., one
school, teacher college, or region), rendering it difcult to make
inferences around scalability. Several larger-scale programme re-
ports/evaluations were identied that did provide valuable insights.
Nonetheless, the lack of grey literature made publicly available im-
plies that there is a signicant amount of (potentially available)
evidence from TPD initiatives that is not informing research and
decision-making processes.
3 EdTech can be successfully leveraged to overcome constraints
operating in LMICs (RQ2 effective tech-mediated TPD). The
literature suggests that technology can be harnessed not only to
support peer learning, but to build bridges across cultures and ge-
ographies, and support the cognitive and attitudinal development of
teachers in LMICs. Moreover, social media, including Facebook and
WhatsApp, have supported teachers within and across countries to
form remote communities of practice, share resources and build
knowledge. Preloaded devices that supply teachers with multimedia
resources to learn from are promising forms of support. Computers
remain the prevalent form of EdTech despite the inux and potential
value of smart devices in increasing reach in LMICs. EdTech has the
potential to provide exible learning methods and modalities
through online, blended learning or self-study opportunities. Self-
study, however, requires that individual teachers have the drive,
professional autonomy and technological skills to engage in profes-
sional learning activities [243].
4 The role of TPD facilitators/coaches emerges as paramount,
although research on how tech can support these pedagogical
leaders is sparse (RQ2). It can provide (semi-)structured observa-
tion tools and prescriptive feedback for use with teachers, for
Exemplar 7
The HALI programme, Kenya
HALI is a health education and literacy intervention that was integrated into teacherseveryday responsibilities [48]. Simplicity was intended in
order to be replicable, using locally sourced and accessed semi-scripted lesson plans and instructional materials along with weekly SMS support.
Teacher engagement remained steady throughout the project. In addition to quantiable gains in knowledge, teachers identied changes in
their responsiveness to student needs (p. 93). A ‘micro-costingor ingredients-based approach was taken, costing each element of TPD support
(e.g., [46]). The three main contributors to cost were: (a) initial training (32%); (b) teacher materials (29%); and (c) SMS support (20%). After
one year of the intervention, teachers knowledge related to early literacy instruction was signicantly higher than that of newcomers to the
intervention. While student learning outcomes were not reported (as the authors acknowledged), they were measured in the subsequent study
by Jukes et al., [80] (Exemplar 1).
16
Both have been published since the review search period.
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
25
instance [185]. Video-recorded observations allow expert coaches or
mentors to connect to the physical classroom quickly and cheaply;
however, they are under-utilised in virtual coaching. Importantly,
despite the common focus on digital technologies in the literature
reviewed, the importance of human relationships was frequently
emphasised. Without relationship building, virtual coaching in
particular can lose efcacy [29].
5 TPD outcomes are mixed (RQ2). Tech-mediated TPD outcomes
include pedagogical and subject content knowledge, support for
teachersmaterial resource creation in low-resource settings, and
increased motivation of both teachers and learners. However, many
challenges arise. The evidence often draws on self-reported data or
fails to identify changes in classroom practices and/or student
learning. Likewise, few studies detailed negative effects of EdTech
within TPD initiatives. While research is often understandably
directed toward measuring its added value to teaching and learning
processes, it is fundamental that any detrimental impact is reported.
This enables the eld to collectively learn and develop understand-
ing of how EdTech can be effectively applied to improve teaching
and learning for all.
6 The ndings reinforce conclusions from previous research that suc-
cessful TPD is designed with and for teachers rather than being
imposed (RQ2). Critical reection through iterative testing in
classroom practice can highlight exactly what contextual adaptations
need to be made to address specic needs [109]. Future research
should examine how using EdTech can more effectively (and ef-
ciently) facilitate the desired deep, critical, ‘structuredreection
[166]. This means building in particular on the emerging implica-
tions for the types and levels of support (both theoretical and prac-
tical) needed to accompany use of videos as stimuli for discussion.
7 Technology-mediated TPD can be used to narrow inequalities
(RQ3), especially through providing greater access to teacher
learning in remote/rural areas, enabling marginalised voices and
enhancing agency. This requires sufcient investment and strategic
planning, though, to avoid inequalities in practice being exacerbated
as observed during the Covid-19 pandemic [44,234]. This reviews
ndings contribute to a more nuanced understanding of the
emerging opportunities and challenges related to equity; nonethe-
less, a large volume of research data were not disaggregated to un-
derstand how initiatives impact certain groups. It is thus difcult to
assess how technology-supported TPD affects different teachers and
learners, and how their needs can be more effectively addressed. For
instance, given the well-documented gender disparities relating to
access and use of technology [240], the tiny number of studies
focusing on TPD to support girlslearning is concerning.
While this review was being conducted, the Covid-19 pandemic led
to an increased use of technology by teachers and learners. However,
support for teachers to adjust to virtual and blended modes of teaching
has been patchy, reinforcing existing inequities: there has not been a
global transformation of how teachers use technology (blog: Wil-
chowski & Cobo, 2021). The needs identied and recommendations
made by this review thus remain highly pertinent.
Recommendations
1 We encourage future TPD programme designers, policymakers, re-
searchers, evaluators, teachers and teacher educators to take ac-
count of multi-level factors across the whole system that
inuence the success of TPD (RQ4); for example, by:
a taking account of teachers professional learning needs, motiva-
tion and agency, in order to increase appropriateness and efcacy
of programmes (micro level);
b working with teachers to co-create TPD models through, for
example, design-based research (micro level);
c ensuring study of the impact of TPD both on teachersknowledge
and skills but most importantly on the subsequent impact on stu-
dentsknowledge and skills (micro level);
d ensuring schools and communities are equipped with the physical
and human resources to support technology-mediated TPD (meso
level);
e developing a deeper understanding of the structural and cultural
factors that can support or constrain technology-mediated TPD
(macro level).
Finally, researchers and stakeholders in this space should also enable
and advocate for greater open access to the evidence, both journal
articles and grey literature, to inform decision-making.
2 From the 170 studies reviewed on tech for TPD in LMICs, these high-
potential evidence gaps emerge:
a more research on larger-scale and longer-term technology-medi-
ated TPD programmes;
b more studies in under-represented countries, particularly by re-
searchers from LMICs;
c strategies for using technology to reach and include marginalised
groups of both teachers and learners, and to cater for diverse
(learning) needs;
d more research on tech-supported TPD in rural settings often
associated with additional challenges (e.g., infrastructure, socio-
economic status, conict/emergency, attendance of girls);
e how the measures undertaken by some researchers to successfully
mitigate the potential detrimental effects of using technology on
social relationships between teachers and coaches/TPD providers
might be applied more widely, including investigating the use of
social media and the feasibility and benets of video-recorded
observations in virtual coaching;
f more research on how tech can be used to support TPD facilitators
and teacher educators (e.g. through scripted coaching software or
virtual learning environments);
g more research on the relationship between technology use and the
levels of structure in pedagogy and lesson scripting that are
appropriate to sustain pedagogical change across contexts this
includes how much structure is necessary in TPD using social
media, and how the effectiveness of informal social media initia-
tives can be maximised.
3 These methodological gaps arise:
a investigating the added value of technology compared to in-person
TPD models;
b measuring cost-effectiveness of initiatives consistently and
comprehensively using experimental methodologies to understand
‘hidden costs;
c undertaking more assessment of impacts on student learning
outcomes;
d conducting follow-up studies to assess sustainability;
e including more stratication by characteristics of teacher partici-
pants (and learners) as well as the disaggregated impact of TPD, e.
g., between teacher groups or across geographical regions;
f strengthening the rigour of reports featuring qualitative data
analysis and validating self-reports;
g conducting reviews of literature published in other languages such
as Spanish, Arabic, French and Chinese.
Author contributions
Sara Hennessy led the team, the conceptualisation and review
writing. Three team members, Sophia DAngelo, Saalim Koomar and
Adam Kreimeia undertook screening, thematic coding and quality
scoring along with writing the conceptual framework and synthesising
ndings. Another core team member, Nora McIntyre, led on database
generation and management, including conducting all of the searches
S. Hennessy et al.
Computers and Education Open 3 (2022) 100080
26
and developing the analytic schemes. Lydia Cao undertook coding and
conducted all of the data extraction, preparing Section 4 and Technical
Appendix F. Meaghan Brugha contributed to coding and quality scoring.
Asma Zubairi analysed and summarised the quantitative data for Section
4.2.
Acknowledgements
We are most grateful to Joke Voogt, Asyia Kazmi, Mary Burns, Dick
Ngambi and anonymous peer reviewers who gave helpful expert input
on earlier drafts of the paper. Thanks are due too for the internal reviews
conducted by EdTech Hub colleagues Katy Jordan, Chris McBurnie,
Louis Major, and Sam Wilson, and to Robbie Carney for assistance with
referencing. The work was supported by EdTech Hub and is funded by
the UK government (Foreign, Commonwealth & Development Ofce,
previously Department for International Development) over the period
20192027.
Supplementary materials
Supplementary material associated with this article can be found, in
the online version, at doi:10.1016/j.caeo.2022.100080.
Appendix A. Historical overview of publications on technology
and teacher professional development
1995 Technology and teacher professional development US
department of educations ofce of educational technology [67]
1999 Technology professional development for teachers [191]
2000 The open learning environment: a new paradigm for inter-
national developments in teacher education [142]
2003 Technology and classroom practices [101]
2005 Using technology to train teachers: appropriate uses of ICT for
teacher professional development in developing countries [53]
2005 Towards a framework for the use of ICT in teacher training in
Africa [228]
2006 DEEP IMPACT: An investigation of the use of information and
communication technologies for teacher education in the global south
[107]
2006 Technological pedagogical content knowledge: a framework
for teacher knowledge [136]
2008 UNESCO ICT competency framework for teachers (version 1)
[223]
2008 International handbook of information technology in primary
and secondary education (Kirschner, P.,