Teaching and learning: what matters for intervention

New learning is crucial for child development. Research in the last century has identified several sub-skills that must come together for the remarkable growth in skills seen during the early years. This research has provided two important insights about child development – a variety of basic processes underpin children’s learning and these bases of learning, while universal, are influenced by the learning context. The study of contexts has spanned multiple levels from home environments and teaching practices to the culture of communities and government policies related to early childhood care and education. More recently, the social, health and educational disruption during the COVID-19 pandemic has emerged as another powerful contextual factor that has deeply impacted learning. In this editorial, I consider five premises to draw attention to the essential core of intervention – the assumption that there will be new learning among children who receive the intervention. The five premises are introduced using examples of typical learning achievements observed in the early childhood years before drawing on papers in this special issue to examine how context and learning interact, and what these might mean for new directions in early childhood intervention research.

Global opportunities and global challenges

The United Nations charter—‘Transforming our World: the 2030 Agenda for Sustainable Development’—has for the first time included as a global goal the task of securing quality learning opportunities in early childhood for all children. Within a blueprint of 17 Sustainable Development Goals (SDGs), Goal 4 is for education and Target 4.2 is to ‘ensure that all girls and boys have access to quality early childhood development, care and pre-primary education so that they are ready for primary education’ (emphasis added). Equally important is an implicit focus on nurturing early child development across all SDGs in the charter. Goal 3, for example, is to ensure healthy lives and promote well-being across all ages and Goal 5 for gender equality and to empower all women and girls. It is clear that the 2030 charter provides an unparalleled opportunity for a global intervention to nurture all children through their early childhood years.

There are, however, several challenges to meeting the SDGs, with 2020 being a watershed year when a pandemic either singly or together with the crises of war and climate emergencies saw hard-won progress roll back (ILO and UNICEF, 2021). In many countries, the losses are in child health and nutrition with more children at risk of abuse and trafficking, and there is a ‘first increase in two decades’ of the numbers pushed into child labour. Thus, before forging ahead to meet ambitious SDG targets, a catch-up to pre-pandemic achievements in child health, nutrition and social protection has emerged as a critical global challenge. Of interest to this Special Issue is the challenge of catch-up in the area of education.

The pandemic led to the shutting down of pre-schools and other early childhood settings, thereby cutting off access to guided learning opportunities outside the home. In addition, when institutions reopened between the waves of the pandemic, the below 6-year-olds were not called back. Forty per cent of the 149 countries surveyed at the end of 2020 reported prioritising older grades, with lower-income countries taking longer to restart pre-primary education (UNESCO, UNICEF and the World Bank, 2020a, 2020b). Thus, for a generation of children primary education will begin with a shaky foundation and is likely to reduce their chances for future success. Catch-up also looks bleak in many countries because the primary school sector, already struggling pre-pandemic, is currently assessed to have made only fair, limited or no progress in ensuring all children complete this level of education (Sustainable Development Goals Progress Chart, 2022, p. 2). Poor primary school education superimposed on disruptions in opportunities for strong foundational learning jeopardises educational attainment in middle and high school, and beyond.

The bases of learning that matter

There are at least five premises about learning that matter for a strong foundation. Here, we use examples from two areas that show rapid growth during early childhood – socio-emotional learning and oral language. Within socio-emotional learning the focus is on emotion recognition in day-to-day social situations that children encounter, and within oral language the focus is on vocabulary learning. Later in the paper the premises are discussed in the context of adult learning.

Premise 1: Learnt skills have deep roots

The first premise about learning is that a ‘narrow’ competence is propped up by a range of processes and skills pulled together for the task at hand (after Railton, 2014). In the area of emotion recognition, for example, several skills support accurate recognition. Take the example of a six-year-old, Child A, learning to accurately recognise the emotions of another child who has lost a toy. Successful emotion recognition implies that several cognitive processes are initiated in parallel or consecutively to read the social context, mentalise about another’s response and find the appropriate vocabulary to label emotion states (e.g. Is this a favourite toy? Is the child unhappy?). In other words, when a child demonstrates a certain learning, the inference is that all the contributing processes necessary for its appearance have developed well. From an interventionist’s point of view, this means that even when a certain skill is a teaching target, it is the component skills that will need the attention. Without this, an essential skill may turn out to have poorly developed contributing skills (hence, shallow roots), reducing the ease with which the learning will be available when needed by the child.

Premise 2: Not all information is relevant for a particular learning task

This is an idea of learning that begins where learning begins – the stream of information available in the surrounding environment. Only some information within any given moment of teaching and learning is relevant for getting a task done or a skill perfected. Learning involves the child knowing which information is relevant and how that information can be usefully included for a task. Here, to learn is to become more sensitive to or selective about what is pertinent in the continuous flow of information surrounding the child. In the just-mentioned lost toy situation, task-irrelevant information may mislead the child’s reading of the social context or another’s experience, and this could result in, for example, naming an emotion with either the wrong valence (e.g. happy) or wrong intensity (e.g. grief).

A similar demand for honing in on relevant information is seen in vocabulary learning. For example, a first-time accurate use of a novel English word such as ‘buzz’ by a four-year-old (Child B) signals that learning of its meaning is well under way and that this learning is based on selecting key units of seemingly discrete data out of a mass of unorganised information. The word ‘buzz’ occurring when bees appear and when the child is physically near a bush of scented flowers are examples of types of linguistic and non-linguistic ambient information that are pertinent to a task. For the learner, the important question that needs resolution is ‘What is the word used to label?’ (e.g. the word ‘buzz’ is used to label the humming sound made by bees). Without the specific realisation of what ambient information to focus on, it is possible that word-to-world linkages will be imperfect and children’s talk will reflect these misunderstandings. Stepping back to learning situations more generally, it is useful to consider whether the learner is overloaded with information and too little is done to help the child select task-relevant information (e.g. when walls are plastered with charts, when there is a continuous stream of spoken instruction).

Premise 3: Learning draws on lower-order rules and higher-order principles

Judgements about incoming information may be based on either rules that are restricted in scope (lower-order) or principles drawn from several circumscribed rules (increasingly higher-order). In other words, to learn is to make inferences about task-relevant information at varying levels of detail and complexity (e.g. Diesendruck & Peretz, 2013). The deepening of information processing may occur at the level of perceptual fine tuning (e.g. noticing more details related to facial expressions, body posture and general demeanour in the lost toy situation) and/or conceptual complexity (e.g. using simple, discrete labels or the more blended labels for emotions). Related to greater sophistication is the ability to abstract and arguably it is this aspect in the learning process that releases the learning for use in novel situations. Transfer of learning to new situations, also called generalisation, can then occur. For Child A, for example, emotion recognition in the lost toy situation could expand to the reading of other social situations related to lost possessions and for Child B, the use of ‘buzz’ could expand to its appropriate use in sentences (e.g. Bees are buzzing outside my window). In the ‘buzzing’ example, the use of a grammatically accurate form demonstrates the learning of a word construction rule in the language (joining the inflection -ing to communicate a particular detail about an event). It is evidence of not just a specific word-to-world mapping but also inferences about the morphological principles for this class of words (thus also invoking premise 1: a learnt skill has deep roots). A similar continuum of deductions will later, perhaps around primary or middle school, underpin the broadening and deepening of vocabulary knowledge when the word ‘buzz’ is also used for a low hum even if there are no bees (e.g. there was a buzz in the room after the announcement). The movement from lower-order inferences to higher-order inferences may be spontaneous for many children but can also be guided through explicit instruction.

Premise 4: Learning draws on domain-general and domain-specific skills

Domain-general skills are those shared across domains of child development such as self-regulation, focussed and sustained attention and planfulness. A powerful domain-general skill is the ability to structure information around regularities and patterns of occurrence. This is a mechanism for implicit learning where children ‘soak up the information around them by mere exposure’ (after Aslin, 2017, p. 1). Domain-specific skills are specialist skills tailored narrowly for the task at hand, such as knowing the right side up when handling a book, recognising letters for word decoding or finding the right way to grip a crayon to make a firm mark on the page. Within the lost toy situation, examples of domain-general processes that may support Child A’s emotion recognition are sensitivity to patterns in incoming information, an attentional bias that makes certain cues salient, and a bias for inductive and/or deductive reasoning. Examples of domain-specific skills for this socio-emotionally demanding task include skills for mentalising about what the other is thinking and feeling, situation appraisal and judging what the situation-emotion co-occurrence may be (e.g. estimating if a lost toy co-occurs more often with feeling upset, fearful or another named emotion). Similarly, an increasingly appropriate use of a word also indicates both general and specialist learning. Domain-general processes that could contribute to word learning include attention and perceptual fine-tuning (noticing the hum because a conversation partner said, ‘Listen!’) and deductive processes (inferring the contexts where ‘buzz’ is used). Domain-general and domain-specific processes are together also necessary to learn when to use the word and when not (e.g. use ‘buzz’ for a soft rather than a loud continuous noise). As with premise 3, the joint use of domain-general and domain-specific skills may be spontaneously well-tuned for many children but can also be guided through explicit instruction.

Premise 5: Experience plays a critical role in learning

This is self-evident but important to consider in relation to emergent skills. If, for example, certain emotion labels are not known to Child A, then the available knowledge-base is limited for certain social situations. Making an efficient judgement about the emotions experienced by the owner of the lost toy will therefore be difficult. If no one notices the child’s silence or provides the child with opportunities to learn by observation or instruction, then there is little opportunity for this specific social-emotional knowledge to grow. That is, when experience with relevant information is itself limited then making considered judgements is compromised. Further learning is yet required.

In the area of vocabulary learning, the phenomenon called referent selection drives the learning about word-to-world and world-to-word connections. It is essential to know what in the world a word is referring to (premise 2 and 3). Helpful labelling of objects, actions or events in the world by a conversation partner such as a sibling, parent, member of the extended family or teacher supports such learning. It follows then that when word usage is inappropriate it is because either word-to-world mapping is underdeveloped in the child or abstraction of the morphological rules of the language is shaky (e.g. as seen in word constructions such as un-buzz, buzz-ly and buzz-ment). In such instances, it will be instruction and the language learning environment that will arguably improve the quality of word representation (e.g. through guided practice, corrective feedback and role models who demonstrate appropriate use of the word). Taking the learning beyond the word as a stand-alone unit to using ‘buzz’ in a constructed sentence will need insights about which other words co-occur with this target word (sentential context, discourse-level context) and in which settings (communicative context). It is expected that multiple and cross-situation exposures will help to incrementally accumulate the learning and consolidate it. These insights go on to serve as the prior experience that is drawn upon in endless new situations where a word may potentially be relevant (e.g. the buzz of mosquitoes on a sultry day, or when drones fly low). It is therefore the opportunity to accumulate and consolidate learning that makes the child’s knowledge base more fit-for-purpose.

Taken together, the five premises offer an inside-out view of what drives learning. The premises also underline the generative aspect of learning and suggest how sensitivity to classes of information develop to reach the status of a dependable knowledge base. While the examples above highlight learning in the cognitive and social-emotional domains, the premises are just as relevant across all domains. For instance, in physical development, another area of rapid learning in the early childhood years, the bases of learning may be for the sub-skills that contribute to movement and locomotion (e.g. run, hop, jump, twirl and somersault) and skills for object use (e.g. throw, catch, skip, pick, kick, climb, write and type). Unpicking the sub-skills for each of these actions helps inform all the different ways in which a comprehensive physical education programme for the early childhood years supports mastery in physical activities. Similarly, the five premises are applicable to other areas that a comprehensive early childhood curriculum typically supports: spatial cognition, mathematical reasoning and music knowledge.

Context and early childhood interventions

Papers in this Special Issue examine interventions from a variety of theoretical lenses using a range of methodological approaches. The theorising that informs the work is for instance about human capital, teacher engagement and language acquisition, with broader theorising from an ecological and cross-cultural perspective. The methods and procedures used include observation, interview, survey, narrative synthesis and paper-pencil tasks; and analytic approaches include basic and more advanced thematic analysis, intermediate as well as multi-level statistical modelling, and triangulation where the research design is mixed methods. The papers cover country settings that vary on wealth and human development indices. The contexts highlight linguistic particularities, belief systems and the political economy of a place, while the interventions range from structured lessons in the classroom to programmes designed to influence teaching practice. Changeable policies and entrenched cultural systems are both examined. All of these are contexts that shape the interventions that are made available to children.

The studies focus on topics included in the rapidly growing area of implementation science: intervention uptake and fade-out effects, scale-up of evidence-informed interventions, monitoring and evaluation, reduction of inequalities, and appropriate research designs for complex social interventions. But the papers also go further by providing analytic insight in several understudied topics: the interactions between teachers’ belief systems and the activities they are ‘recommended’ to deliver (Jukes et al., 2022), the design elements in teacher materials when scripted interventions cross borders to enter new settings (Ramacciotti et al., 2022), the components in an intervention when the children’s spoken language is distinct and different from the language variety encountered in kindergartens (Saiegh-Haddad, 2022), the potential of ‘direct assessment tools’ to balance the current reliance on reported data from parents or teachers when making multi-country comparisons (Richards et al., 2022), the effects of home disadvantage on children’s ‘learning trajectories’ when resources at home differ (Zhao et al., 2022), and what is to be considered as ‘impact’ for an ambitious multi-tiered family-focussed intervention and what is its ‘legacy’ when policies change (Sammons et al., 2022). These papers therefore cover single-domain interventions (mainly language interventions) and multi-domain (the early childhood curricula in its entirety). Some are clearly theoretically driven and, importantly, use different lenses for a grounded analysis (psycholinguistics: Saiegh-Haddad, 2022; cultural: Jukes et al., 2022; the political economy: Zhao et al., 2022). The analytic unit in these interventions includes child outcomes, teacher outcomes, intervention components, service users and non-users, and government policies.

Several cross-cutting themes emerge. These include the role of influential actors in children’s learning including parents, teachers, the educational bureaucracy and policymakers. Other themes include responding to children from vulnerable groups, acculturative efforts and challenges to implementation fidelity. Also embedded across papers is how context must support the design of the next generation of interventions: designing learning programmes informed by local social goals and language complexities (Jukes et al., 2022; Saiegh-Haddad, 2022); localising teacher professional development through ‘scripted suggestions’ and child assessment, especially for emergent literacy and child wellbeing (Ramacciotti et al., 2022; Richards et al., 2022); and broadening the barometer of efficiency when evaluating early childhood policies: is it a ‘start stop’ policy landscape and do attainments of the most disadvantaged high achieving 5-year olds slip one year on (Sammons et al., 2022; Zhao et al., 2022). These topics make explicit the processes that have stayed invisible in intervention work and together underline the importance of a nuanced understanding of the contexts within which early years interventions are embedded.

The five premises revisited

We return to the five premises about the bases of learning to further examine implications for intervention design and the uptake of interventions at all levels of the education system, especially the teacher. Premises 1 and 5, a learnt skill has deep roots and experience plays a critical role in learning, have been discussed above in terms of underpinning cognitive processes, and with a focus on child learning. These premises apply just as much to the cultural and interpersonal roots to skill development, and to teachers learning new skills. In a case study from rural Tanzania, even when teachers were highly recommended that they follow the constructivist, individual-focused, ‘active approach’, adoption was low (Jukes et al., 2022). The reticence was for activities such as giving children individual practice and calling on them to explain or answer targeted questions. One explanation for the teachers’ pedagogical decisions was their interpersonal sensitivity; teachers did not want to single out a child or cause embarrassment. They ‘generally did not agree with the rationale for differentiated instruction’. Instead, choral reading was a preferred practice in this teacher community, a finding recorded in several countries (Nag et al., 2016), as was learning by working together.

Another area of challenge in teacher training is reported in a process report on material development and a pilot in a private school in urban Brazil (Ramacciotti et al., 2022). The teaching activities that remained most resistant to new learning among participating teachers were a) for sub-skills that underpin visible learning (for vocabulary learning: ‘make deeper semantic connections and not just name’) and b) to consolidate learning (through ‘extension and expansion’, the ‘closing off activities’). An intervention in preschools in Israel for Palestinian-Arabic-speaking 5-year-olds, also with a focus on language learning but within the more linguistically complex situation of dialects and diglossia, used the scripted lessons to especially focus on ‘continuously highlighting differences and similarities’ between the dialect and the standard varieties (Saiegh-Haddad, 2022). This later was expected to aid the development of meta-cognition. The idea of meta-cognition is particularly of interest in any discussion on learning because it simultaneously invokes a shift to higher-order inferences and greater specialist skills in a specific domain (premises 3 and 4, Learning draws on lower-order rules and higher-order principles and domain-general and domain-specific skills). Finally, an analysis of learning gains in preschools from three states of India finds that early high-achievers, who do not have the advantage of strong home resources, fall behind as early as age 6. The authors consider what role a ‘play-based’ curriculum with more ‘flexible classroom organisation’ may have in explaining a less dramatic fall in the states that use these approaches to teaching and learning.

The findings from the four studies bring to the fore three important tensions in the field of teacher training: first, that new learning in adult learners will confront established skills with deep cultural roots (premise 1), and yet too often professional development for teachers has proceeded without engaging with teachers’ beliefs, values and preferences; second, some of the most difficult areas for teacher professional development are related to deepening and fine-tuning learning (premises 2, 3 and 4); and third, that new learning needs exposure, time for practice, support that is clear and comprehensive, and with ‘real examples’ (premises 4 and 5). The tensions appear to be around ideas of quality in early childhood education. From the point of view of intervention developers, a focus on quality translates into enforcing fidelity, an adherence to what is prescribed. However, this is a technicist approach and also suffers from the assumption that the same set of teacher skills are sufficient and effective in all contexts. Instead, quality is to start by listening to teachers’ perspectives and aiming to ensure, especially for teacher-focused interventions, that the intervention has relevance, is comprehensive and focussed on higher-order principles (rather than lower-order rules) that allow teachers to intentionally adapt and be responsive to different teaching-learning situations. This appears important to gain acceptance and for sustainability but most importantly for directly improving the quality of the early childhood education for the child. Implicit in all of this is the need for urgent attention to under-researched teaching practices found to be preferred in different contexts as well as the innovations that have emerged to address unique characteristics of a given context. Without research into local pedagogical preferences and local innovations, the global evidence base for early childhood interventions will remain askew.

The call to attention in this Special Issue of three research tools is relevant. First, for the study of cultural influences, Jukes et al. (2022) call for improved methods to elicit and examine teacher beliefs and their explanations for teaching decisions. Their study demonstrates the power of a vignette-based method for eliciting and analysis. Second, Sammons et al. (2022) call for a realist approach to mixed-methods designs alongside theoretically driven statistical modelling. Their narrative synthesis of nearly two decades of policy interventions and their evaluations demonstrate the potential of this approach to address complex system-level processes, especially large policy interventions that inevitably include ‘naturally occurring variations in patterns of use’ of early childhood provisions. Finally, I have delineated in the current paper the bases of learning that matter and demonstrated their use as a critical lens to qualitatively evaluate interventions. Advances in the field of early childhood intervention research are likely to gain from all these tools.

Acknowledgments

This Special Issue has gained enormously from the peer reviewers who gave generously of their time and expertise. They are Gideon Arulmani, Yonas Asfaha, Leon Feinstein, Charles Hulme, Sandra Mathers, Joshua McGrane, Diane Mayer, Portia Padilla, Benjamin Piper, Marina Puglisi, Margaret Snowling and Shaher Banu Vagh. I would like to acknowledge Annette Riziki’s assistance and Maria Evangelou, who first mooted the idea of a Special Issue. Sonali Nag’s editorial work and this paper were supported by the ‘UKRI GCRF Supporting Children’s Oral Language Development’ award. The support of the Economic and Social Research Council (ESRC) is gratefully acknowledged.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Notes on contributors

Sonali Nag

Sonali Nag’s research seeks to develop a nuanced, contextually grounded understanding of child development. She has worked extensively on the languages of south Asia and how children master the scripts of the region. Her work on the effects of contextual factors on literacy learning is based primarily on an unparalleled database of research in low– and middle-income countries. This work has, for instance, drawn attention to large-scale replication of western tests even when education is not in a European language or the alphabetic system. Similarly, these reviews highlight how potentially useful cultural practices linked to oral and choral language traditions and learning–by–writing are missing in early literacy interventions. These lines of study have relevance for policy and practice, and for the next generation of programmes promoted by multilateral agencies.