JCPSLP Vol 17 No 3 2015

Table 1. Summary of appraised articles, key findings, strength of evidence and levels of evidence

Strength of evidence (Millar et al., 2006)

Key findings

Source/title of evidence

Level of evidence (JBI, 2014) and critical appraisal checklist used Level 1 (Meta-analysis) EVIDAAC rating scale for systematic reviews (Schlosser et al., 2008) Level 2 (Randomised controlled trial) JBI MAStARI: RCTs or non- RCT group intervention studies appraisal checklist (JBI, 2014) Level 2 (Pre-test post-test quasi- experimental/ observational cohort study) JBI MAStARI: cohort or case control studies appraisal checklist (JBI, 2014) Level 3 (Longitudinal cohort study) JBI MAStARI: cohort or case control studies appraisal checklist (JBI, 2014)

Does discovery-based instruction enhance learning? (Alfieri et al., 2011)

Preponderant 1

Unassisted discovery learning methods do not benefit learners, whereas enhanced discovery learning methods (including feedback and scaffolded play) were found to enhance learning outcomes. Child centred play therapy (CCPT) can significantly increase academic achievement for 1st grade academically at-risk children (compared to waitlist control group). A play-based curriculum was more effective than a traditionally structured classroom curriculum in increasing children’s play, social skills, and narrative language.

Play therapy in elementary schools: A best practice for improving academic achievement. (Blanco & Ray, 2011) Play, language and social skills of children attending a play- based curriculum school and a traditionally structured classroom curriculum school in low socioeconomic areas. (Reynolds et al., 2011) End-of-phase 2, Report 4: Outcomes for pupils over time . (Sproule et al., 2009)

Preponderant

Preponderant

Preponderant

Children participating in a play-based curriculum had more positive attitudes towards learning than children in traditional classrooms. No long-term differences in educational outcomes.

1 Preponderant: study had minor flaws with respect to: design, dependent variable reliability OR treatment integrity, e.g., “outcomes were more likely than not to have occurred as a result of the intervention” (Millar et al., 2006, p. 251)

Sarahlouise White (top) and Sue McAllister

problem-solving, maths, science, and computer skills. A large number of studies were included in the analysis and both fixed and random effects statistical models yielded similar findings. This gives confidence in the finding that guided play-based learning was more effective than didactic or free play. Specifically, children achieved better educational outcomes when engaged in learning via scaffolded play that involved teacher support and encouraged learners to explain their ideas (elicited explanation). This style of learning yielded better outcomes than any other form of instruction (free play or explicit instruction), and led to superior results for all learning outcomes including language (verbal and social tasks). However, caution is required when interpreting the findings with regard to children in early primary education, given that the authors did not assess the quality of the included studies, and included studies pertaining to learners of all ages (43% of included studies related to child population). The randomised controlled trial by Blanco and Ray (2011) compared academic achievement outcomes of 41 first grade children (aged 6–7 years) in the United States who were randomly assigned to either a 2 x ½ hour play therapy group or a waiting list control group. During sessions, the facilitator followed the child’s lead and responded to the child’s play following a protocol of verbal and non-verbal behaviours intended to communicate connection and caring. Although the study did not specifically evaluate play-based learning in an educational setting, this article was included due to the relevance of its findings regarding enhancement of educational outcomes including spoken language, reading, writing, and maths as a direct result of play, as measured by the Young Children’s Achievement

Eight studies met the inclusion criteria. Four articles were selected for individual critical appraisal on the basis of their relevance to the clinical question, level of evidence and recency (see Table 1). The level and strength of evidence based on methodological rigour was assessed independently by the first two authors using standardised critical appraisal checklists tailored to speech pathology and/or allied health evidence appropriate to each study type (Joanna Briggs Institute, 2014; Schlosser et al., 2008). An additional four studies (Miller & Almon, 2009; Van Oers & Duijkers, 2012; Walsh, McGuinness, Sproule, & Trew, 2010; Walsh, Sproule, McGuinness, & Trew, 2011), although relevant to our question, were excluded on the basis of their comparatively lower levels of evidence. The strength of evidence for each article was rated using Millar, Light, and Schlosser’s (2006) levels of certainty. Each appraised article was rated as “preponderant” due to minor flaws with respect to design, dependent variable reliability, or treatment integrity (Millar et al., 2006). This strength of evidence indicates that the outcomes were more likely than not to have occurred as a result of the intervention (Millar et al., 2006). The following is a summary of the critical appraisal for each of the four selected articles. Findings The meta-analysis by Alfieri, Brooks, Aldrich, and Tenenbaum (2011) directly compared two levels of play (“unassisted discovery-learning” [free play] and “enhanced discovery-learning” [guided play]) with explicit/didactic teaching methods. This comparison provided useful insight into different types of play-based learning and their effects on educational outcomes including verbal/social skills,

156

JCPSLP Volume 17, Number 3 2015

Journal of Clinical Practice in Speech-Language Pathology