Table 6

Results of the mixed-effects ANOVAs for the qualitative moderator variables on the effect sizes obtained from measures of knowledge in the posttest
Cluster / Moderator variable k d+ 95% C. I. ANOVA results
dl du
(A) Treatment characteristics:
Type of treatment: QB(2) = 1.08, p = .583
Postural hygiene (PH) 12 1.301 0.821 1.781 R2 = 0.0
PH + Physiotherapy exercise (PE) 3 1.521 0.562 2.479 QW(13) = 359.20, p < .001
PH + Physical activity 1 0.537 −1.060 2.134
Type of postural hygiene: QB(2) = 3.67, p = .159
Acquisition of knowledge (AK) 6 1.394 0.810 1.979 R2 = 0.251
AK + Posture training habits (PTH) 8 1.432 0.954 1.911 QW(13) = 210.58, p < .001
AK + PTH + Stimulat. dynamic postures 2 0.449 −0.472 1.369
Teaching method of postural hygiene: QB(1) = 0.18, p = .669
Theoretical teaching (TT) 6 1.407 0.736 2.077 R2 = 0.0
TT + Practical teaching 10 1.225 0.730 1.720 QW(14) = 341.28, p < .001
External agents: QB(1) = 1.63, p = .201
Yes 10 1.119 0.694 1.545 R2 = 0.244
No 6 1.592 1.005 2.178 QW(14) = 268.85, p < .001
Parents as paraprofessionals: QB(1) = 0.76, p = .384
Parental involvement 1 1.972 0.385 3.559 R2 = 0.003
No parental involvement 15 1.244 0.842 1.647 QW(14) = 377.45, p < .001
Teachers as paraprofessionals: QB(1) = 6.77, p = .009
Teacher involvement 5 0.730 0.231 1.230 R2 = 0.497
No teacher involvement 11 1.544 1.189 1.898 QW(14) = 173.17, p < .001
Mode of application: QB(2) = 11.15, p = .004
Indirect intervention 1 0.289 −0.793 1.371 R2 = 0.481
Direct intervention 12 1.578 1.234 1.923 QW(13) = 144.57, p < .001
Mixed intervention 3 0.534 −0.118 1.185
(B) Methodological characteristics:
Type of control group: QB(1) = 1.65, p = .198
Active control 2 1.972 0.860 3.083 R2 = 0.028
Nonactive control 14 1.194 0.783 1.605 QW(14) = 367.82, p < .001
Evaluator blinding: QB(1) = 0.35, p = .553
Blinded evaluator 10 1.372 0.902 1.841 R2 = 0.078
Not blinded evaluator (or not reported) 6 1.133 0.503 1.764 QW(14) = 305.94, p < .001

k: number of studies. d+: mean effect size. dl and du: lower and upper confidence limits of the 95% confidence interval around the mean effect size. QB: statistic for testing the significance of the moderator variable. QW: statistic for testing the model misspecification. R2: proportion of variance accounted for by the moderator variable.

Calvo-Muñoz et al.

Calvo-Muñoz et al. BMC Musculoskeletal Disorders 2012 13:152   doi:10.1186/1471-2474-13-152

Open Data