Building the Foundation of Aquatic Literacy in 4–6 Years-Old Children: A Systematic Review of Good Pedagogical Practices for Children and Parents

Children between the ages of 4–6 years represent the population most affected by drowning accidents, while their early involvement in physical activity, and more specifically in aquatic activities is a key factor in their future physical life journey. The systematic review’s purpose was to identify aspects in the intervention’s studies with children and/or their parents that had a significant impact on the Aquatic Literacy (AL) dimensions mentioned as motor, psychological, affective, and cognitive. The PICO method was used to define the research question and PRISMA checklist searched for articles in nine databases: Cochrane, Embase, ERIC, ProQuest, PsychInfo, PubMed, Scopus, SportDiscus, and Web of Science. Eligibility criteria were: (1) English language, (2) primary research, (3) population of 4–6 year old children or their parents, (4) intervention study design, and (5) results related to at least one of the AL domains. The strength of evidence and the risk of bias were assessed. Results showed relatively poor number of studies for such a vulnerable population regarding the drowning risk (n = 8 for parents and n = 14 for children intervention). Studies did not show a consensus on which educational approach was more beneficial than others. Concerning parental education, results were rather homogeneous, especially concerning the theoretical frameworks employed and the relevancy to include parents in swimming programs. The development of pedagogical tools for promotion and evaluation, based on the AL theoretical framework, could help to clarify the question of “how to teach” children to prevent drowning and engage young children in long-term physical activities.


Introduction
Words such as "diving", "swimming", "floating", and "gliding" are terms evoking the pleasure humans can experience from aquatic pursuits. Indeed, it comes as no surprise that different forms of aquatic recreation figure prominently amongst our favorite pastimes [1,2]. In view of the sedentary lifestyle and the lack of physical activity (PA) observed for many years among young people [3], water can be a rich and abundant medium for PA, allowing a sustainable commitment to an active and healthy lifestyle [4]. Aquatic activities are often promoted for their health benefits [5][6][7][8], and positive childhood experiences can have a significant impact on lifelong engagement with them [9], encouraging educators to support maximum enjoyment and confidence in children in these environments. However, these benefits can be tragically overshadowed if children are not empowered against the possibility of drowning. This "dark side" of water recreation is partly responsible for the deaths of more than 365,000 people worldwide every year, half of whom are children and in theoretical framing was highlighted by Stallman [28] who proposed the concept of "Water Competence" to bypass the epistemological debate about the definition of "can swim" and to focus on "what needs to be taught" during swimming lessons. This heterogeneity of theories underlying existing research is reflected in the multiplicity of concepts used. A first level of conceptualisation is anchored exclusively in the dimension of motor skills. There is a noticeable frequency of typical notions in the scientific and pedagogical literature such as "swimming ability", "swimming skills", "aquatic skills", "survival skills", "rudimentary skills", "water readiness", "pre-requisite skill level", "swimming proficiency" [29]. This terminology is focused on behavioural indicators regarding what needs to be learned in order to move safely in water. A second level of concepts is used, such as "drownproof", "drowning prevention" and especially the concept of "Water Competence" initiated by Langendorfer and Bruya [26], and, more recently, by Stallman et al. [28]. At this more holistic level of conceptualisation, "Water Competence" highlights an ability to cope with a potential risk of drowning by mobilising motor capacities, but also cognitive (making decisions, knowledge about the aquatic environment and humans around) and affective resources (managing emotions). Stallman et al. [28] identified 15 water competences that could significantly reduce the risk of drowning. The importance of each of these competences was supported by research findings, addressing the question: "what needs to be taught to protect children from drowning". This proposal is now widely accepted and has inspired many training programs around the world [30][31][32]. A third level of conceptualisation is inspired by the 'physical literacy' framework [33] which is defined by "the motivation, confidence, physical competence, knowledge and understanding that an individual possesses and enables him to value and take ownership of his commitment to physical activity throughout their lives" [34]. Physical literacy for aquatic environments or 'Aquatic Literacy' (AL) appears very close to the concept of water competence, including positive motivations to engage in water activities [35].
According to Denehy et al. [36], the lack of theoretical framing in aquatic competence research has inhibited the development of a clear understanding of good practice and its underpinnings in physical education. In a meta-analysis, Leavy et al. [37] confirmed this idea by showing that the majority of prevention programs and their assessment have been implemented without a theoretical framework to frame and understand the behaviours of young children or their parents. This deficit makes the overall message difficult to identify. Moreover, if the questions of "optimal age" and "what to teach" have been widely studied by the researchers, the "how to teach" question tends to be based on pedagogical beliefs, rather than evidence-based interventions [38]. Therefore, it is essential to identify the most effective intervention modalities for the most vulnerable population (young children and their parents). This focus will not only help to guide early "how to swim" teaching strategies to develop participant safety, but also would ensure engagement in future aquatic activities throughout the lifespan, forming the foundation for their aquatic literacy.
In this paper, we present an up-to-date systematic analysis of the effectiveness of teaching strategies used for young children (4-6 years old), and their parents, based on empirical data. The aim, designs and methodological quality of the intervention studies are presented and discussed critically. The effects of the interventions are examined on different dimensions (motor, psychological, social, and cognitive). With the information given in this systematic review, we aim to inform and support stakeholders in the field of aquatic education, from curriculum makers to pedagogues, in order to contribute to the overall quality of aquatic education.

Materials and Methods
This systematic review was carried out by using the protocol developed by Cochrane Institute [39] and following the checklist of the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines [40]. The aim of this systematic review was to identify whether there are evidence-based methods to optimise the development of aquatic literacy in young children.

Literature Identification
Based on the PICO protocol (Participants-Intervention-Comparison-Outcomes), the research question has been formulated (i.e., For pre-school children aged 4-6 years with no specific health conditions and their parents, what characteristics of educational interventions are effective in improving safety and/or engagement in, on and around the water as compared to children from the same age range who did not benefit from those interventions?) and relevant keywords were selected and are presented in Supplementary Data S1.
The keywords related to each of the following categories have been added using the boolean operator "OR" and search techniques, such as truncation and/or phrase marks, adapting them to each database: population (e.g., preschool * OR parent *), intervention (e.g., "swim * training" OR "aquatic lesson *"), outcomes (e.g., "tread * water" OR pleasure), generic terms (e.g., drown * OR "aquatic literacy"). These categories were then combined using the boolean operator "AND" and some terms related to certain pathologies were excluded using the "NOT" operator (e.g., depression OR asthma) to obtain the search algorithm. This algorithm was run during between 28 October and 15 November, 2021, on nine online databases: Cochrane, Embase, ERIC, ProQuest, PsychInfo, PubMed, Scopus, SportDiscus, and Web of Science. The specific search strategy for each database explored are presented in Supplementary Data S1.

Selection Process
After collecting the results from databases, duplicates were removed. Then, titles and abstracts of each article were screened independently by two researchers (LM and CS). Disagreements on study eligibility were resolved by an external reviewer (FP). Reference lists of studies were also screened for potentially eligible records.
To be included in the full-text analysis, studies were required to satisfy the following criteria: (1) English language, (2) primary research, (3) part of the population between 4-and 6-year old children or their parents, (4) intervention study design, and (5) results related to at least one of motor, psychological, social or cognitive aspects.
The exclusion criteria of studies were: (1) children with disabilities or pathologies, (2) articles which the 4-6-year-old population has not been subdivided, and (3) articles without specifying the age of the children.

Data Extraction
Relevant information has been extracted from the full-text analysis and then moved to an Excel spreadsheet. The following information were: (1) bibliographic information (title, authors, year of publication), (2) strength of evidence [41], (3) objective, (4) theoretical framework(s), (5) intervention characteristics, (6) sample size and characteristics, (7) measurement tool(s), (8) outcomes (referring to the four physical literacy domains of Keegan et al. [42]), (9) quantitative and qualitative summary according to group or subgroup when available, and (10) risk of bias.
The data were extracted independently by two researchers and an external opinion was solicited from a third researcher (FP) in case of disagreement.
Studies were categorised according to their measurements towards children exclusively, children and their parents, and parents exclusively.

Level of Confidence and Risk of Bias Assessment
Two tools were used to assess the studies quality: (1) the strength of the evidence grid of Ackley et al. [41] and (2) risk of bias grid adapted from the Cochrane checklist [39].
To assess selection bias, we examined whether the populations were randomised and the percentage of dropouts (<20%). For information bias, we assessed whether the group was comparable at baseline characteristics, if the baseline values were accounted for, and if the intervention were blinded for population and examiners. Finally, for the bias analysis, we assessed whether the timing of measurement was comparable between intervention and control groups, adequate statistical procedure and presence of p-value, effect size and confidence interval. Two researchers (LM and CS) independently evaluated the study quality according to these two tools. Disagreements on study quality were resolved by an external reviewer (FP).

Overview
The PRISMA flowchart in Figure 1 shows the process of identification, screening, and inclusion of studies. The search yielded 976 results. After removing duplicates, a total of 838 titles were screened. Of these, 725 were excluded during the title screening, and 54 during the abstract screening phase, leaving 50 studies for full-text assessment. During this stage, 12 more articles were included in the full-text analysis by a snowballing process and screening the bibliography and references from the relevant articles. the percentage of dropouts (<20%). For information bias, we assessed whether the group was comparable at baseline characteristics, if the baseline values were accounted for, and if the intervention were blinded for population and examiners. Finally, for the bias analy sis, we assessed whether the timing of measurement was comparable between interven tion and control groups, adequate statistical procedure and presence of p-value, effect size and confidence interval. Two researchers (LM and CS) independently evaluated the study quality according to these two tools. Disagreements on study quality were resolved by an external reviewer (FP).

Overview
The PRISMA flowchart in Figure 1 shows the process of identification, screening, and inclusion of studies. The search yielded 976 results. After removing duplicates, a total o 838 titles were screened. Of these, 725 were excluded during the title screening, and 54 during the abstract screening phase, leaving 50 studies for full-text assessment. During this stage, 12 more articles were included in the full-text analysis by a snowballing proces and screening the bibliography and references from the relevant articles. Twenty-one studies were selected for this review. Table 1 (1 study [43]), Table 2 (13 studies [44][45][46][47][48][49][50][51][52][53][54][55][56]), and Table 3 (7 studies [22,[57][58][59][60][61][62]) show the main characteristics of the interventions in terms of objective, level of confidence, theoretical framework used, population, intervention description, measurement tools, and outcomes domains for studies including children and their parents, only children and only parents, respectively. Figure 2 shows a summary of the AL domains concerned in the outcomes of the studies, and Figure 3 shows the risk of bias analysis.  Twenty-one studies were selected for this review. Table 1 (1 study [43]), Tab studies [44][45][46][47][48][49][50][51][52][53][54][55][56]), and Table 3 (7 studies [22,[57][58][59][60][61][62]) show the main characteristic interventions in terms of objective, level of confidence, theoretical framework us ulation, intervention description, measurement tools, and outcomes domains for including children and their parents, only children and only parents, respectively 2 shows a summary of the AL domains concerned in the outcomes of the stud Figure 3 shows the risk of bias analysis.

Main Characteristics of Interventions towards Parents
Seven studies concerned parents exclusively (Table 3), and one study concerned parents associated with children ( Table 1). The samples built for the design studies ranged between 15 [59] and 4010 [60] parents. The duration of the programs is variable: single campaigns [58] up to a 6-week campaign [60], instructions or feedback about children during their swimming teaching ranging from 10 weeks of teaching [62] to 8 months of teaching [22], and training ranging from 20 min [59] to 1 h [61]. Five studies were explicitly anchored in a theoretical framework [43,[58][59][60][61]. Intervention methods were assessed by measuring effects of different communicational and pedagogical variables: drowning prevention campaign [58], presence during the children swimming lessons [22,43], effect of close-call drowning [57], receiving feedback about children improvement [57], parental self-instructional programs [59], and parental water safety programs [60][61][62].
Over the eight studies explored, only three used a control group [57,60,61] to strengthen the evidence analysis. Only two studies used validated questionnaires in their experimental design [59,61], four studies used questionnaires created specifically for the study [22,43,57,58,62], and one study used video analysis [60].

Outcomes Explored through the Aquatic Literacy Concept
The outcomes measured in children analysis concerned the motor (n = 9), psychological (n = 5), cognitive (n = 5), and social (n = 2) dimensions. Only two studies covered three dimensions: motor, psychological, and social [43,55]. The types of effects measured in the parent studies were cognitive (n = 7), psychological (n = 3), and social (n = 1). Figure 2 synthesizes the domains explored by each studies reviewed.

Discussion
The aim of this systematic review was to identify pedagogical evidence to: (1) improve pedagogical strategies for building the foundations of early childhood aquatic education, and (2), identify intervention strategies for parents to raise awareness and improve their attitudes towards the risk of drowning. The main results showed that: (1) although the literature displays the existence of different pedagogies, capable of improving different dimensions of aquatic literacy, deciding how to teach aquatic skills depends on the teacher's pedagogical objectives; and (2), parents should be included in the educational programs to maximize drowning prevention.

A Relatively Poor Number of Studies for Such a Vulnerable Population Regarding the Risk of Drowning
Our systematic review of nine international databases showed a limited number of studies that addressed the question of how to teach children to swim in the age range 4 to 6 years. Many epidemiological studies throughout the world have highlighted the vulnerability of this population to drowning [10], providing a legitimate educational question from a scientific perspective. The plethora of learning programs aimed at this age group in different institutional settings confirms Stallman's point [38] that the pedagogies proposed for use are based more on beliefs than on scientific evidence. While the questions of "when and what" to teach have been deeply explored in the scientific literature, the question of pedagogical methods (how to teach) needs further exploration. This weakness in the applied scientific literature can, in part, be explained by the methodological and epistemological issues involved in research in the contexts of real-life teaching and learning. Carreiro Da Costa [86] highlighted an evolution in the theoretical frameworks used in research on effects of pedagogies used, initially anchored in a positivist process-product paradigm (measuring the effects of one or more variables, while neutralising the effects of others). The impossibility of identifying obvious regularities in samples of teachers and students within educational settings has led researchers in the field of educational interventions to adopt mixed methods approaches. These include seeking evidence of the teacher's thinking and actions and observing and recording multiple interactions with, and between, students. This complexity of measuring the teaching-learning process in an aquatic environment for children may have limited research projects in this area. Across the 21 studies included in our review, the lack of an explicit theoretical framework (n = 8) or the variety of frameworks employed relating to communicating science [58], public health [43,50,60,61], motor learning [43,47,49], psychological development [44,52,55] and water skills acquisition [45,52], corroborate the conclusions of Denehy et al. [36] and Leavy et al. [37]. For them, the heterogeneity, or even the absence of theoretical frameworks, make it difficult to summarise the main findings in order to develop clear and consensual recommendations.

What Are the Criteria for Selecting One Pedagogical Approach, Rather Than Another?
The challenge addressed by our PICO methodology was to identify the pedagogical characteristics that significantly impact the progress of students and their parents, in limiting risks of drowning, while encouraging children to engage in aquatic physical activity throughout life to develop aquatic literacy. Analysis of all the studies focusing on effects of pedagogical characteristics of aquatic programs revealed a high level of heterogeneity in theoretical frameworks implemented measurement tools and data used to assess children's learning outcomes. Most of the intervention studies we found emphasised a skill-or competence-based approach and helped to question the effectiveness of different pedagogy types on the physical dimension of physical literacy [42]. The results highlighted three main findings: In terms of the temporal and quantitative characteristics of the programs, Erbaugh [47] compared the evolution of children's swimming skills exposed to programs of different duration. They reported that the length of the exposure to interventions was the main driver of swimming skill acquisition. Bradley et al. [49] confirmed this trend by showing similar progress in comparing a massed vs. distributed swimming teaching programme.
Concerning skill acquisition, the use of environmental constraints (shallow vs. deep water learning, use of video feedback) showed a significant effect on skill acquisition and/or improvements in confidence [45,52].
The type of pedagogy adopted also provides interesting insights. To exemplify, although most studies proposed the use of competence-based and top-down structured interventions [50], some pedagogies, based on less structured activities and play also improved water-related physical competences [44].
In that regard, the pedagogy selected for use seems to rely on the priorities of the teacher and their intentions to improve the one or more dimensions of aquatic literacy. Invernizzi et al. [43] analysed the effect of two types of pedagogy: one more structured (linear pedagogy, or LP) and another which emphasised less structured learning (nonlinear pedagogy, or NLP). Their results showed that LP improved the reproduction of more specific aquatic skills compared to NLP, but that the latter simultaneously increased the confidence and enjoyment of learners more, stimulating further learning, than LP. Other interventional studies evidenced a combined effect of environmental constraints on combined dimensions of aquatic literacy. Costa et al. [45] showed that shallow water teaching led to the learning of more advanced motor skills, since children felt more confident, which encouraged them to explore new motor possibilities. They highlighted methodological biases by considering that, in deep water, the program could not be delivered in the same way for safety reasons (using more directed pedagogy, and occasional use of flotation devices). Rocha et al. [52] reported similar findings using a similar methodology, with improvements emerging in a broader spectrum of water skills when teaching in shallow water. However, the results showed that these between-group differences in skill acquisition disappeared after 18 months of instruction. These results, which favour teaching in shallow water, are the only ones permitting comparisons of similar outcomes. However, since those effects vanished over time, and the potential biases due to the use of flotation equipment, shows that both approaches (shallow vs. deep water teaching) could be equally relevant according to the learning context (short or long exposure). Arhesa et al. [44], without using a control group, showed that a 'play' based pedagogy improved water skills and confidence in young children. Bunker et al. [54] assessed the effects of video feedback on front crawl learning and showed significant effects in flutter skills kick performance compared to the control group.
We also found evidence that the dimensions of physical literacy could be improved, whether related to the cognitive [46,48,51,53,56], psychological [53], or social dimensions [43,55]. Overall, these studies showed significant improvements in all domains investigated and allowed the authors to validate the putative benefits of their respective programs. Taken together, those results showed that many dimensions of physical literacy might be subject to improvement, so it is up to the teacher to define and prioritise their objectives and select a method accordingly.
However, the overall quality of the study design (poor to high risk of biases), the lack of control-randomised studies and the multiplicity of pedagogical characteristics, and studies with heterogeneous methodologies call for caution in interpreting and applying the results in practice. The data highlight that more research on the topic of aquatic literacy development is needed.

What Kind of Interventions Addressed to Parents Help Improving Safety and Protection for Their Child?
Our selection of articles highlighted the importance of implementing specific programs to address parental needs, as children's participation in learning to swim programs might paradoxically expose them to additional risks of drowning.
Several studies have highlighted parental underestimation of the risks of drowning, an overestimation of their children's ability to cope with these risks, and unsafe activity monitoring attitudes towards their children [57,60,61]. Morongliello et al. [22] highlighted that the more parents perceived their children's progress as successful, the lower their perception of the risk of drowning, and the weaker their supervision attitudes when their offspring were near the water. These results corroborated previous work [57] showing that when parents were regularly updated on their children's progress, they developed less of a monitoring attitude. However, parents with near-drowning experiences remained more vigilant, regardless (with or without progress feedback). Strategies to tackle this phenomenon are diverse, ranging from involving the parents in the teaching process, to the implementation of communication campaigns, information seminars, and provision of training for parents in first aid, independently from the teaching process. Results suggested that expanding "can swim" programs to enhance parents' perceptions and understanding of children's drowning risks and supervision needs, significantly mitigates protection against drowning. Delivering messaging in the form of 'close-call' drowning stories are especially effective to impact sustainable parental supervisory practices [57].
Different interventional strategies have been conducted in the literature to test their efficacy by measuring their impact on beliefs, knowledge, drowning risk, monitoring be-haviours, and first aid. The first type of intervention is the use of traditional communication tools (poster, signage, information cards and fact sheets) to expose parents to information and gain their attention and interest on risks of drowning for their children [58,60]. Quan et al. [58] showed positive effects in terms of raising parents' awareness. Other programs based on this intervention type have been inconclusive. For example, the "Keep Watch @ Public Pools" program showed that parents of children aged 0-5 years had relatively low supervision scores for attention (2.6/4), proximity to children (2.2/4), and readiness (2.8/4). This programme had no effect on these behaviours [60]. There seems to be converging evidence on the importance of including parents in learning to swim schemes in order to make them aware of drowning, as motor progress shown by children may have adverse effects. The second type of intervention consists of a more parent-focused program about water safety knowledge and attitudes [61,62] and a self-instructional first-aid program about knowledge and confidence to give Cardio-Pulmonary Resurrection [59]. Studies of direct interventions on parents (seminars, courses during children's lessons, or first aid training) have shown positive results in terms of improving attitudes about child supervision [59][60][61][62].
Few studies have specifically examined interventions with parents (n = 7), although the results appear to be more homogeneous than similar studies with children. This is chiefly due to a greater homogeneity in the measurement tools used, and the variables (knowledge, beliefs, supervision) and theoretical frameworks mobilised (communication sciences and behavioural change sciences). Although the levels of evidence are still relatively low, the results are consistent in highlighting the importance of including parents in swimming classes and making them aware of the need for increased supervision, especially when children are making significant progress in aquatic skills.

The Need for a Common Theoretical Framework and Tools to Promote and Assess the Aquatic Literacy of Young Children
The diversity of theoretical frameworks and measurement tools used in previous studies makes it difficult to summarise the results, particularly regarding the most relevant educational variables for teaching young children to swim. The question of "how to teach" young children, therefore, appears to be limited, not only by their quantity, but also from an epistemological point of view. While the concept of 'water competence' [28] has become one of the frameworks to guide programs in the question of 'what to teach', our results suggest that studies simultaneously addressing the effects of teaching interventions on motor, cognitive, psychological, and social aspects remain scarce. However, protecting children by equipping them with relevant motor skills, knowledge, and attitudes, necessary to face the dangers of drowning, while instilling the pleasure of moving in an aquatic environment, appears to be the major pedagogical challenge with regard to mitigating the tendencies towards a sedentary lifestyle in many young people [87].
The concept of aquatic literacy seems to be necessary to unify knowledge on the question of "how to teach" children of all ages. This concept, emerging from the broader concept of physical literacy, could provide the same impetus for researchers to create measurement tools and identify benchmarks to identify strengths and weaknesses in current programs. Recent studies in the field of physical literacy have shown very promising developments in the implementation of programs to educate young people to adopt an active, safe, and sustainable lifestyle [88]. The recent publication of the winners of the Erasmus Sport + program suggests that tools and benchmarks will soon be available to guide aquatic educational programs in the same way. The "ALFAC" (Aquatic Literacy for All Children) program aims to create an international database through the cooperation of researchers in seven countries (France, Lithuania, Germany, Portugal, Poland, Norway, and Belgium), based on the creation of a test battery to measure aquatic literacy. Similar to the PISA program, these benchmarks and new tools could be used as a guide to develop aquatic education programs which will allow researchers to develop common tools to facilitate the comparison and synthesis of their results.

Conclusions
The purpose of this systematic review of the literature was to examine the current state of research in the field of aquatic education in order to identify evidence-based pedagogical features to improve the early education of young children and their parents.
Our results highlighted the need to formally include parents in this process. The findings also exemplified programs which successfully developed physical, psychological/emotional, social, and cognitive dimensions of behaviour, empowering teachers with the capacity to diversify among different outcomes according to learners' needs. However, due to the lack of intervention studies, the medium-to high-risk of bias, and the fact that none of the selected articles addressed all four dimensions of aquatic literacy simultaneously, the question of "how to teach aquatic literacy" remains open to further investigation.
Science has two aspects. One is knowledge that has been sufficiently validated and confronted with empirical experience to be considered reliable. The questions of "what to teach" and "when to teach" seem to fall into this category. The other aspect of science is that of research, that of questions that have not yet found a robust response. The question of "how to teach" is part of this issue. We hope that there will be a mobilisation of the theoretical framework of aquatic literacy, as well as development of the tools that will be produced to educate future generations, both on the dangers of water, but also on the sheer enjoyment and pleasure to be experienced in this specific activity context.