The experimental group outperformed controls on only one of five budget questions — identifying maximum stock levels from an inflow–outflow graph — with no significant advantage on the others¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r1 · q2
Currency · newest study 2015 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? A proportional odds model showed a significant positive treatment effect for Question 4 (coefficient 1.13, p = 0.007, adjusted p = 0.035), while Questions 1–3 were nonsignificant and Question 5 did not warrant a model. → Reichert et al. 2015 q2 i? Question 4 correct responses rose from 6% on the pretest to 22% on the final (p < 0.001), whereas Questions 2, 3, and 5 showed no statistically significant increase. → Reichert et al. 2015 (2)

Evidence¶

Reichert et al. 2015¶

Collin Reichert, Cinzia Cervato, Dale Niederhauser, and Michael D. Larsen. (2015). Understanding Atmospheric Carbon Budgets: Teaching Students Conservation of Mass. Journal of Geoscience Education 63, 222–232. https://doi.org/10.5408/14-055.1

q2 · i? · causal · r1

Proportional log odds ratio models (Table IV) fit to per-question change scores for four budget questions; Question 4's coefficient was 1.13 (p = 0.007, adjusted p = 0.035) and Question 5 was not modeled due to few changes. Two alternative methods gave consistent conclusions.

"Questions 1–3 have nonsignificant results. Question 4 has a significant positive impact before and after adjustment for multiple testing."

Reichert et al. 2015 (2)¶

Collin Reichert, Cinzia Cervato, Dale Niederhauser, and Michael D. Larsen. (2015). Understanding Atmospheric Carbon Budgets: Teaching Students Conservation of Mass. Journal of Geoscience Education 63, 222–232. https://doi.org/10.5408/14-055.1

q2 · i? · causal · r1

McNemar's test for correlated proportions on pretest-to-final percentages for all five questions; Questions 2 (p = 0.18), 3 (p = 0.58), and 5 (p = 0.68) did not show statistically significant increases.

"Question 1 (pretest 48%, final 86%, p < 0.001) and Question 4 (pretest 6%, final 22%, p < 0.001) showed statistically significant increases in the percentage of correct answers."

Discussion¶

Related Claims¶

• Pupils taught with the learning technologies reached a higher complex assessment of educational skills than control pupils (63.4% vs 50.4% of maximum) — related
• Reliance on the pattern-matching misconception dropped more in the experimental group than in the control group but persisted for the majority of students — related
• The experimental group outperformed the control group on every one of the five visualization tasks individually — reports the opposite