Meta Analyses

Summary of Play Therapy Meta-Analyses Findings

Four meta-analytic studies examining play therapy effectiveness have been conducted since 2000. Leblanc and Ritchie (2001) published the first play therapy meta-analysis using hierarchical linear modeling (HLM) to explore the treatment effect of play therapy among 42 controlled outcome studies (number of child participants not reported) completed between 1947 and 1996. Bratton, Ray, Rhine, and Jones (2005) conducted a meta-analytic review with a random effects model for 93 humanistic and non-humanistic (behavioral) play therapy studies (3,248 child participants in total) completed between 1953 and 2000. Lin and Bratton (2015) employed HLM techniques to explore the effectiveness of child-centered play therapy (CCPT) from 52 controlled outcome studies (totally 1,848 child participants) completed between 1995 and 2010. Ray, Armstrong, Balkin, and Jayne (2015) reviewed 23 controlled studies (totally 1,106 child participants) completed between 1975 and 2011 with a random effects model to evaluate effectiveness of CCPT conducted in elementary school settings.

Meta-Analyses

# of Studies

Study Collection Years

Child Mean Age (years)

# of Participants

Effect Size

Leblanc & Ritchie (2001)

42

1947-1996

7.8

Not reported

0.66

Bratton et al. (2005)

93

1953-2000

7.0

3,248

0.80

Lin & Bratton (2015)

52

1995-2010

6.7

1,848

0.47

Ray et al. (2015)

23

1975-2011

Range: 4-13

1,106

0.21-0.38


Overall Treatment Effect

Bratton et al. (2005) reported a statistically significant overall effect size of 0.80 for the 93 studies, indicating a large treatment effect for play therapy. Similarly, Leblanc and Ritchie (2001) reported a statistically significant overall effect size of 0.66 for the 42 studies, indicating a medium to large treatment effect for play therapy. Based on the enormous total sample size, these meta-analytic findings provided robust research evidence for the treatment effect of play therapy.

Lin and Bratton (2015) explored contemporary controlled studies employing CCPT interventions and estimated a statistically significant overall effect size of 0.47 for the 52 collected studies, indicating that CCPT demonstrated a medium treatment effect. Ray et al. (2015) calculated effect sizes by target outcomes instead of an overall effect size for school-based play therapy. Ray and colleagues found statistically significant effect sizes in the small to medium range for total problem behaviors (ES = 0.34), internalizing problem behaviors (ES = 0.21), externalizing problem behaviors (ES = 0.34), self-efficacy (ES = 0.29), and academic achievement (ES = 0.36).

Specific Presenting Problems/Issues

Bratton et al. (2005) reported statistically significant effect sizes for children’s behavioral problems (ES = 0.81), social adjustment (ES = 0.83), personality concerns (ES = 0.80), self-concept (ES = 0.51), anxiety/fear (ES = 0.69), developmental/adaptive concerns (ES = 0.90), and parent child relationship (ES = 1.12), indicating play therapy is an effective treatment for a variety of children’s presenting issues and parent child relationship stress.

Lin and Bratton’s (2015) findings confirmed CCPT’s effectiveness for a variety of presenting issues or problems. The statistically significant effect sizes they reported include global behavior problems (ES = 0.48), internalizing behaviors (ES = 0.42), externalizing behaviors (ES = 0.33), self-efficacy (ES = 0.63), academic achievement (ES = 0.46), and parent child relationship stress (ES = 0.59). Ray et al. (2015) provided further support for the effects of CCPT on specific outcomes including total problem behaviors (ES = 0.34), internalizing problem behaviors (ES = 0.21), externalizing problem behaviors (ES = 0.34), self-efficacy (ES = 0.29), and academic achievement (ES = 0.36).


Treatment Setting

According to the statistically significant effect sizes reported by Bratton et al. (2005), play therapy is considered effective in a variety of settings, including school settings (ES = 0.69), outpatient clinics (ES = 0.81), residential settings (ES = 1.10), and critical-incident settings (ES = 1.00). Ray et al. (2015) further confirmed CCPT’s effectiveness in school settings for a variety of presenting problems.

Treatment Format

Bratton et al. (2005) reported similar statistically significant effect sizes for group play therapy (ES = 0.73; 33 studies) and individual play therapy (ES = 0.70; 34 studies), supporting the efficacy of play therapy in both individual or group formats.

Treatment Duration

According to Leblanc and Ritchie (2001) and Bratton et al. (2005), the optimal play therapy treatment effect falls within 30 – 40 sessions when the treatment is provided by mental health professionals.

Treatment Fidelity

Lin and Bratton (2015) reported statistically significantly larger effect size (ES = 0.58) for studies complying with rigorous treatment protocols (such as the use of treatment manuals, clear treatment procedure, and appropriate therapist training) compared with the studies with lower level of treatment fidelity (ES = 0.21). This finding supports the importance of treatment fidelity in CCPT treatment process.

Parent Involvement

Leblanc and Ritchie (2001), Bratton et al. (2005), and Lin and Bratton (2015) reported statistically significant differences between studies with full parent involvement and studies with partial to no parent involvement, although the mean effect sizes for both groups (partial to no parent involvement and full parent involvement) were statistically significant. Ray et al. (2015) did not include play therapy studies with parent involvement.

Meta-analysis

Full Parent Involvement

Partial or No Parent Involvement

Leblanc & Ritchie (2001)

ES = 0.89*

ES = 0.56

Bratton et al. (2005)

ES = 1.15*

ES = 0.72

Lin & Bratton (2015)

ES = 0.59*

ES = 0.33

Note. * indicates a statistically significance between the two effect sizes.

Meta-analytic findings confirm the importance of parent involvement in children’s treatment process and outcome.


Child Characteristics

Age. Lin and Bratton (2015) reported statistically significant mean effect sizes for studies with an average age of 7 years and younger (ES = 0.53; 42 studies) and studies with an average age of 8 years and older (ES = 0.21; 10 studies), with a statistically significant difference between these two effect sizes. This finding suggested that children of all ages can benefit from CCPT, especially children younger than 8 years of age.

Ethnicity. The findings of Lin and Bratton (2015) revealed statistically significant mean effect sizes of 0.76 for 15 CCPT studies mainly recruiting non-Caucasian children and 0.33 for 15 CCPT studies mainly recruiting Caucasian children, with a statistically significant difference between the two groups. This finding suggested that CCPT is a culturally responsive counseling intervention for diverse child populations.

Sex. Leblanc and Ritchie (2001), Bratton et al. (2005), and Lin and Bratton (2015) reported sex of child as a non-significant predictor for treatment effect, which suggested that play therapy is equally effective for males and females.


Conclusion

Meta-analysis is considered useful for overcoming the typical weakness of small sample size often found in individual studies investigating treatment effectiveness with clients (Lipsey & Wilson, 2001; Whiston & Li, 2011). Play therapy meta-analytic findings clearly confirm the effectiveness of play therapy, specifically CCPT approaches including filial therapy. It should be noted that the effect sizes reported in these four meta-analytic reviews are Cohen’s d effect sizes, which can be interpreted as “how many standard deviations of difference between the mean scores of two groups.” For example, the effect size of 0.5 means the experimental group performed 0.5 standard deviations higher than the control or comparison group.

Note: The relatively smaller effect sizes reported by the two recent meta-analytic studies targeting CCPT approaches cannot be interpreted as lower treatment effect for CCPT compared to earlier meta-analyses. Plausible explanations for the smaller, yet noteworthy effect sizes may include the use of more rigorous research methodologies in contemporary outcome studies and differences in meta-analytic techniques employed in each study.


References

Bratton, S., Ray, D., Rhine, T., & Jones, L. (2005). The efficacy of play therapy with children: A meta-analytic review of treatment outcomes. Professional Psychology: Research and Practice, 36(4), 367-390. doi:10.1037/0735-7028.36.4.376

LeBlanc, M., & Ritchie, M. (2001). A meta-analysis of play therapy outcomes. Counseling Psychology Quarterly, 14, 149-163. doi:10.1080/09515070110059142

Lin, Y., & Bratton, S. C. (2015). A meta-analytic review of child-centered play therapy approaches. Journal of Counseling and Development, 93(1), 45-58. doi: 10.1002/j.1556-6676.2015.00180.x

Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: Sage.

Ray, D. C., Armstrong, S. A., Balkin, R. S., & Jayne, K. M. (2015). Child-centered play therapy in the schools: Review and meta-analysis. Psychology in the Schools, 52(2), 107-123. doi: 10.1002/pits.21798

Whiston, S. C., & Li, P. (2011). Meta-analysis: A systematic method for synthesizing counseling research. Journal of Counseling and Development, 89(3), 273-281.