Improving children’s language, literacy and mental health: Evaluating the impact of the classroom promotion of oral language (CPOL) approach.

In 2014, teachers in the intervention arm of the study will attend three face-to-face days of professional learning convened by the research team, and will also engage in a self-directed manner in an online learning network of teachers in like-schools in the first half of 2014. A fourth day of face-to-face learning will be held in February 2015. The professional learning delivered to teachers is based on Munro’s (2011) ICPALER (Ideas, Conventions, Purposes, Ability to Learn, and Reception) Framework, and was the subject of a successful large-scale pilot led by CI Snow (Snow et al., in press). In addition to these formal days of professional learning and access to the online forum, teachers will liaise with CPOL Support Officers, via intermittent face-to-face, telephone, and online contact, in order that questions are addressed and program fidelity is enhanced.

2014-04-28

2015-12-31

Primary outcome:
When the study cohort is in Grade 3 (2017), we wish to evaluate student literacy skills, and pla to access and use their National Assessment Program Literacy and Numeracy (NAPLAN) data as one outcome measure against which to evaluate the impact of the CPOL investigation. NAPLAN is made up of tests in the following four domains:
• Reading
• Writing
• Language Conventions (spelling, grammar and punctuation)
• Numeracy.
A full description of the way in which NAPLAN is administered is available at: http://www.nap.edu.au/naplan/naplan.html NAPLAN is administered by the Australian Curriculum Assessment and Reporting Authority (ACARA) - http://www.acara.edu.au/default.asp

Primary outcomes:
Child Literacy at Grade 3 (2017) , assessed through students’ grade 3 NAPLAN score: ‘Reading’ category

Secondary outcomes:
• Child oral language at Grade 3 (2017), assessed through:
• Students’ NAPLAN scores: Categories including: ‘Writing’, ‘Language Conventions’, and ‘Numeracy’
• Reading Progress Test: End Grade 1 (2015)
• Additional language measures: CELF-4 subscale, Bus Story Test, NPVT (digitised version of the Peabody Picture Vocab Test) End Grade 1 (2015)
• Child mental health end Grade 1 (2015), assessed through: Strengths and Difficulties Questionnaire
• Teacher practice assessed at the end of Prep (2014) and end Grade 1 (2015) (Teacher oral recording/Self reflection/Teacher Survey/Teacher focus groups)

Cluster Randomised Controlled Trial

Computer randomisation.

Participants are clustered at the school level:

Schools will be selected for inclusion in this study based upon the 2009 and 2012 Australian Early Development Index (AEDI) data for their school. Schools with AEDI results in the language and cognitive skills domain with vulnerability rates of ≥ 10% who had at least 10 children participate in the AEDI will be eligible to participate in this trial. DEECD and CECV will arrange their own calls for Expressions of Interest from the pool of eligible schools across the state of Victoria. From this pool, schools that were identified as vulnerable in both the 2009 and 2012 AEDI data will be approached first, after which schools vulnerable only on the 2012 AEDI will be approached.

All children in one classroom at each school will be allocated together as a single unit for randomisation (‘the cluster’). This is the most appropriate approach to randomisation given the intervention will be focused on teacher professional learning and the implementation of that knowledge within a whole-of-classroom approach. It will also limit the potential for contamination by separating the trial arms at the school level. Following baseline data collection, schools will be stratified by sector and allocated randomly - using a random number table - to one arm by an independent, blinded statistician otherwise unconnected with the study.

Yes

72 schools

1450 participants including students and teachers.

36 schools control, 36 schools treatment

The sample size calculation is bound by the assumption that 80 points is the average gain in the NAPLAN Reading score over 2 years, therefore around 40 is the average gain in the Reading score over 1 year. For this study we aim to find a difference between the intervention and control groups in the Reading score of 0.3 SDs (22.47 points). An effect sizes of 0.3 SDs can be meaningful at a population level given the reach. In this trial in particular it is an approximate 6 month difference in progress and would represent a “clinically” significant difference in outcomes. Randomisation of 561 children per arm is required to provide 90% power to detect a minimum difference of 0.3 standard deviations on the NAPLAN ‘Reading’ scores at grade 3, allowing for an average intra-class correlation coefficient of 0.08 and an average cluster size of 17. To allow for a potential attrition rate of 20% of children by the time children are in grade 3, 700 children per arm (1400 in total) will be required in the study. This sample size also allows for a retention rate of 8% of clusters.