MIT-based instruction shows a large overall positive effect on students' academic achievement in a meta-analysis of 75 Turkish experimental theses and dissertations¶

Claim · All claims
Evidence · 1 study (2 entries) · quant-synthesis r2 · q2 · i3 large
Currency · newest study 2016 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i3 Random effects meta-analysis yields an overall effect size of d = 1.077 (95% CI 0.921–1.232), large by Cohen's classification, significantly favoring MIT-based instruction. → Baş 2016 q2 i? Effect sizes across the 75 studies were heterogeneous (Q(74) = 458.793, p < .001), so a random effects model was adopted. → Baş 2016 (2)

Evidence¶

Baş 2016¶

Baş, G. (2016). The effect of multiple intelligences theory-based education on academic achievement: A meta-analytic review. Educational Sciences: Theory & Practice, 16, 1833–1864. https://doi.org/10.12738/estp.2016.6.0015

q2 · i3 · quant-synthesis · r2

Random effects synthesis of 75 experimental theses and dissertations on MIT-based instruction in Turkey, coding pretest/posttest means and SDs into Hedges' g. The effect was "d = 1.077", significant by Z-test, and large under Cohen's (1992) classification.

"When the total number of 75 studies included in the meta-analysis were compared according to the random effects model, the standard error was found to be 0.079 and the upper limit in the 95% confidence interval was 1.232, while the lower limit was 0.921 with an effect size of d = 1.077"

Baş 2016 (2)¶

Baş, G. (2016). The effect of multiple intelligences theory-based education on academic achievement: A meta-analytic review. Educational Sciences: Theory & Practice, 16, 1833–1864. https://doi.org/10.12738/estp.2016.6.0015

q2 · i? · quant-synthesis · r2

Cochran's Q heterogeneity test on the pooled effect sizes. The variance "is larger than can be explained by simple sampling error", so the random effects model was adopted rather than the fixed effects model.

"This Q value indicates that the distribution of effect sizes in this collection of studies was heterogeneous, Q (74) = 458.793, p < .001"

Discussion¶

Related Claims¶

• Nearly all (97.33%) of the included studies show effect sizes in favor of MIT-based instruction over control-group instruction — related
• Computer-based teaching has a large, positive, significant effect on students' academic success compared with traditional teaching methods in Turkish studies from 2006-2014 — related
• The 5E instructional model shows a large, significant pooled effect on academic achievement versus traditional teaching methods in Turkish studies (2008-2014) — related
• Computer-Assisted Mathematics Education has a large positive overall effect on academic achievement (ES = .8999) in Turkish experimental and quasi-experimental studies — related
• Instruction tailored to students' learning styles shows a large pooled random-effects effect on academic achievement (d = 1.029) across 29 Turkish experimental studies — related
• Context-based learning has a statistically significant positive overall effect on academic achievement (g=0.970) compared to traditional lecture-based instruction — related
• The constructivist learning approach has a strong positive overall effect on students' academic achievement compared to traditional teaching methods (ES = 1.156). — related
• Control groups taught with traditional methods in Turkish geometry and numbers theses show a large average pretest-posttest gain of about 0.83 standard deviations — related
• Effect of MIT-based instruction does not differ significantly between master's theses and doctoral dissertations — related