Design

The project will use a balanced-clustered RCT design to follow one cohort of children from kindergarten programme entry (2020) and into the first year of school (Preparatory; 2021).

Setting

In Queensland, Australia, the year prior to formal school commencement is known as kindergarten. Participation rates in this year are over 95% nationally25 and children attend either a long day care centre, with the potential to attend up to 12 hours per day 5 days per week (as enrolled by their family), or a stand-alone kindergarten programme that typically offers either a 5-day fortnight programme (6 hours per day) or a 2-day per week programme with longer days (7.5 hours per day). In stand-alone kindergartens, the same cohort of children is enrolled in each programme, as opposed to long day care where different groups and combinations of children attend on any given day of the week. The national curriculum for the years prior to school, the Early Years Learning Framework,26 guides the play-based approach in both types of programmes. All kindergarten programmes employ a bachelor-qualified teacher and a teaching assistant as well as other inclusion support educators as needed. The adult to child ratio is 1:11, with stand-alone kindergartens generally enrolling up to 22 children per group. Children are typically 3.5–4.5 years at commencement of the kindergarten year and enrol in kindergarten for 12 months.

We will engage with one major early childhood education and care service provider (Childcare & Kindergarten Association; C&K) to support recruitment of kindergartens. C&K, established in 1907, is a Queensland based, large, not-for-profit organisation with 350 kindergartens and childcare centres. C&K manages two types of services—branches and affiliates. Branches are where C&K is the licensee of the service and responsible for the management of staff, workplace issues, insurances, information technology, legislative compliance, buildings and maintenance and funding.27 Affiliate centres have entered into a service agreement with C&K, where C&K distributes funding, provides support and audits compliance obligations on behalf of the Queensland Government.28 C&K is providing no financial or other resources to the project and is not being paid for their collaboration in centre recruitment. A formal research approval process with C&K has been completed.

Centres to be recruited will meet the following criteria: affiliate or branch member with C&K; located within 70 km from Brisbane, Australia; be in a low socioeconomic area as indicated by a score in the bottom three deciles of a national index based on census data that position areas in Australia according to relative socioeconomic advantage and disadvantage29; be in a community area where 2018 Australian Early Development Census data indicate a higher than national average level of child vulnerability in both domains of social competence and emotional maturity30; run a kindergarten programme with the same cohort of children attending at least 2 consecutive days per week (ie, stand-alone kindergarten rather than long day care); have a teacher in the kindergarten programme who is not planning to retire, resign, transfer or take long service leave during the trial period (last quarter 2019–2020); have a typical start-of-year enrolment of at least 80% of capacity in each group (17 children in a group with capacity to take 22).

Kindergarten teacher recruitment

Directors at all eligible centres will be emailed trial information by C&K and invited to complete an expression of interest form, which will collect information further checking on eligibility items (including anticipated teacher employment stability). From those centres who complete an expression of interest, an independent researcher not involved in the study will randomly select the study centres, and teachers within the study centres, using stratification of child enrolment capacity, provided by teachers on the expression of interest form. Some teachers will teach up to 22 children, if they teach in one programme, and up to 44 children, if they teach in two programmes, across the week. If multiple teachers from one centre return an expression of interest, only one teacher from each centre will be randomly selected to reduce the introduction of another level in analyses. The independent researcher will select an even number of teachers with 22 children and teachers with 44 children to ensure a balanced design across two conditions is possible. As well as sociodemographic and context measures (table 1), all teachers will complete a baseline survey collecting information on outcome measures of frequency of music, rhythm and movement practices; confidence in using rhythm and movement; level of knowledge on self-regulation and music and movement for brain development; beliefs about the role of active music engagement for development; training and experience in education and specifically in self-regulation and rhythm and movement (table 2).

Randomisation

Randomisation will occur at the centre level (which is also the teacher level) following recruitment of centres and the teachers within the centre and collection of baseline teacher data. An independent researcher will use a random number generator again stratifying by teacher-provided child enrolment numbers to ensure a balanced number of children in both intervention and control groups. This is because teachers will either teach 22 children or 44 children across a given week depending on their centre or work arrangements as described above. Stratification will mean an equal number of smaller units and larger units will be assigned to each condition to achieve a balanced design. Intervention teachers will then train in programme delivery, practicing with their 2019 preschool cohort. In 2020, a new cohort of preschool children will be enrolled in the centres and invited to participate in the research, meaning randomisation must occur prior to child recruitment and baseline measures.

Intervention training

Teachers in the intervention centres will be provided with intervention materials and professional development in the form of brief self-paced online orientation covering the rationale for the programme; full-day workshop covering self-regulation in early childhood, the theory and rationale for use of rhythm and movement and exposure to all of the intervention materials, microteaching demonstrations of the intervention (teachers act as children so they can experience the activities themselves); a demonstration led by the research team in their centre; independent practice of sessions with their 2019 class; observation and coaching by the research team; and engagement with the online professional learning community for tip sharing and trouble shooting. Prior to the implementation period, an online refresher course will be completed. In 2021, following the completion of data collection, all control teachers will gain access to RAMSR materials and training.

The intervention

RAMSR is designed to be delivered by early childhood practitioners in a group setting with up to 25 children. Teachers will deliver this as part of their usual group floor or circle time within the kindergarten daily routine. Typically, this is held in the morning as a warm up to the day. The RAMSR programme consists of a series of four session plans, with each plan to be repeated daily for 2 weeks. In these settings, because of enrolment patterns, this will mean each child receives the programme four to five times per fortnight, yielding a total of 16–20 sessions available to each child over the 8-week intervention period.

All activities within the session plans are designed to practice key skills of attentional, emotional and behavioural regulation, inhibition, shifting and working memory through embedding these skills in coordinated movement activities enhanced by rhythmic auditory cueing. Common activity elements across the sessions included start/stop (inhibition), reversal of instruction (shifting, for example, move in the silence and freeze in the music), working memory games and beat synchronisation to changing tempos. Original backing tracks provided rhythmic auditory cueing and leverage rhythmic entrainment principles to stimulate more coordinated movement.

Within each session plan, there are a series of activities with the above key elements represented in each (1) warm-up involving body percussion, (2) becoming familiar involving an adaptation of a familiar early childhood song, (3) moving to the beat involving large gross motor movements, (4) playing to the beat involving simple rhythm sticks or castanets, (5) dancing to the beat involving slightly more complex gross motor movement patterns to activity three of the session and often involving visual-motor skills and coordination such as mirroring the shape of rhythm sticks on the floor with bodies and (6) calming, which includes a series of movements and then stillness accompanied by relaxation music to support physiological entrainment to a calmer state, targeting embodied emotional regulation. Teachers are provided with each session plan, a series of videos demonstrating the plans, all audio tracks required to run the activities and visual cue cards needed for some activities. Rhythm sticks and animal castanets required to implement the sessions are provided. Further and regularly updated information on RAMSR, including training approaches and opportunities, can be found at https://ramsrblog.wordpress.com/.

Child recruitment and baseline measures

Teachers will invite all children enrolled and commencing the 2020 kindergarten year between 28 January 2020 and 21 February 2020 (first 4 weeks of the kindergarten year) to participate in the research through gaining parental consent. Validated baseline self-regulation measures (parent and teacher report and direct assessments with children) will be taken with all children prior to the intervention/control period (baseline; see table 3). Assessors will visit kindergartens and withdraw children to undertake assessments and will be blind to group assignment. Assessors will visit centres in pairs, with each child working with each assessor for a 15-minute session, splitting the tasks into two sessions. Kindergarten teachers will provide checklist measures of children and will not be blind to group assignment.

Intervention delivery

In term 1, teachers will undertake a brief refresher training seminar (2 hours) prior to commencing delivery of the programme 2-3 times weekly for 8 weeks. Teachers will access ongoing support through regular phone contact and access to the closed online professional learning community. Once every 2 weeks, a research team member will visit to observe the teacher delivering the session, to take fidelity measures and to provide coaching. Coaching will follow protocol by Hemmeter et al.31 Each postsession reflection will involve: open, summarise observation and encourage reflection; supportive performance feedback; corrective performance feedback; targeted support; support for planned actions; end meeting. A summary of each verbal coaching session will be provided via email to teachers and this will become a source of research data to address our second research question. Teachers (and coaches when observing) will provide fidelity checklist information following the completion of each session, which will collect data on: adherence to the session plan; level of child engagement in sessional activities as a measure of participant responsiveness; attendance data of children to determine dosage. The research team lead will also make additional visits to all pairs of teachers and coaches to collect additional independent fidelity checklist information. Programme differentiation across the control and intervention groups will be ascertained through examination of baseline and follow-up data collected on teacher surveys regarding the frequency of use of rhythmic movement sessions in kindergartens.

Follow-up

Follow-up data collection will occur at the end of the 8-week intervention period, and 12 months after baseline, after children have transitioned to school (table 2). All participating families will be telephoned early in 2021, when most children will be enrolled in Prep (first year of full-time schooling in Queensland). Parents will be asked to provide contact details for their child’s school and current teacher and to confirm consent for the research team to contact the school. Prep teachers will then be asked to complete the self-regulation measure for each child (paper form delivered to the school), providing follow-up data for children following school transition. School teachers will be blind to group assignment. Teacher report has been selected as the only data collection approach at the 12-month follow-up period, rather than a repeat of direct child assessments undertaken at baseline and 8-week follow-up. This is because it is anticipated that study children will attend up to 50 different schools spread across a wide geographic region and it is not feasible within financial constraints of the study to undertake direct child assessments. However, as depicted in table 3, four of the teacher-report measures are repeated across all three time points, which will allow for some consistency in the analyses.

Measures

The measures used throughout the study are summarised in tables 1–3. Measures have been selected with reference to the known psychometric properties and feasibility of measures in low socioeconomic kindergarten settings, many of which were used in the pilot study for the current trial.22 The project lead will provide training for face-to-face measures to the data collection team. The training will involve familiarisation with the measure’s components and how they are administered. Mock assessments will be conducted to further familiarise the staff in the presence of the trainer. The lead researcher will also observe each assessor once in the first week of data collection, and ongoing troubleshooting meetings will be held to ensure fidelity.

All electronic and paper data will be collected against a unique digital identifier for each participant. Data will be directly uploaded to a database for the three iPad tasks. The Design Copy task will be double scored (blind) by the lead author and a single research assistant, with inter-rater reliability calculated. For the two other child assessment measures, data will be collected through data collection paper forms. Data entry will be managed by a single data entry research assistant with a 10% random selection of forms for each measure double checked by the lead author. Confidentiality is ensured through storage of all data against numeric identifier codes only. The identification key will be stored separately and only accessible by the lead researcher and senior research assistant.

Sample size

Children: pilot study estimates have been used for power and sample size calculations. Assuming a simple sampling design with no clustering, calculation in G*Power indicates a required sample size of 76 to detect small intervention effects (0.14) with a power of 0.8 using the planned analytic approach (path analysis equivalent to multiple regression). The balanced cluster design of the RCT produces a design effect of up to 1.9 (using average intra-class correlation coefficients for same measures used in the pilot of 0.03 and average cluster size of 33), increasing the effective sample size to 144. Using the 2016 pilot response rates at similar centres as a guide, a sufficient final sample of at least 144 at the 12 month follow-up time point will be achieved by recruiting a minimum of 200 children at baseline (allowing for a 15% attrition rate at each of the two follow-up time points).

Teachers: with only up to eight teachers involved in the study, split across intervention and control, the study is underpowered to address hypothesis C using significance testing. Instead, we will use a match case-series approach and descriptive statistics to take an exploratory approach to understanding any baseline similarities and differences in the teacher constructs among intervention teachers and matched control teachers and change scores in constructs across baseline to follow-up.

Statistical analyses

The efficacy of the intervention will be evaluated within a structural equation modelling framework (using MPlus) with an intention-to-treat approach, which will allow for: precision in the testing of intervention effects on the outcome measures across the intervention and control groups while controlling for baseline scores and covariates; handling of missing data using full information maximum likelihood and accounting for the clustered design.32 To reduce the number of analyses undertaken, multiple measures for each construct will be reduced to single scores using appropriate approaches for each as shown in table 3. For primary outcomes (executive function and self-regulation) and the secondary outcomes measured equivalently at baseline and 8 weeks later (school readiness and visual-motor integration), four separate path analyses (equivalent to multiple regression) will be used with effect sizes computed using formulas for independent groups pre–post designs.33 To examine 12-month follow-up data, latent curve modelling for pre–post-post designs will also be used to examine intervention effects across baseline to postintervention to follow-up34 for self-regulation factor scores that are equivalently measured at three time points. For the secondary outcomes measured only at 12-month follow-up (academic competence, behaviour problems and school transition), path models examining the effect of treatment condition on these outcomes, controlling for baseline measures and mediated by 8- week follow-up measures will be developed. Clustering will be accounted for at the centre level. Covariates will be selected from child gender and age, level of family income, parental education, non-English speaking status and aboriginal status. Selection will be based on consistency and significance of bivariate correlations among these covariates and outcomes measures.

The first phase of analysis will be undertaken once baseline and postintervention data collection is complete. A second phase of analysis will be completed once the 12-month follow-up data collection is complete.

Participant and public involvement

Participants and the public were not involved in the design or reporting of this research protocol and will not be involved in the design, reporting and dissemination of the trial.