Low prior knowledge students learned more from PyKinetic's Parsons problems than high prior knowledge students, especially in the experimental group with SE prompts¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q3 · i3 large
Currency · newest study 2019 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i3 LP students in the experimental group had a Cohen's d of 1.95, more than double the LP control group's d of .91, supporting that Parsons problems with incomplete LOCs are more beneficial for low prior knowledge learners. → Fabic 2019 q3 i? LP students in the experimental group had significantly higher normalized gains than HP students for all questions (U = 66, p = .024), procedural questions (U = 68, p = .014), and easier-to-guess questions (U = 68.5, p = .011). → Fabic 2019

Evidence¶

Fabic 2019¶

Fabic, G. V. F., Mitrovic, A., & Neshatian, K. (2019). Evaluation of Parsons Problems with Menu-Based Self-Explanation Prompts in a Mobile Python Tutor. International Journal of Artificial Intelligence in Education, 29, 507–535. https://doi.org/10.1007/s40593-019-00184-0

q3 · i3 · causal · r2

Analysis of the top and bottom 25% of participants by pre-test score (9 LP and 9 HP in experimental; 7 LP and 7 HP in control). Table 8 prints LP experimental d = 1.95 vs. LP control d = .91; HP experimental d = −.31 vs. HP control d = .50. The authors state this shows "evidence of H3."

"there is a substantial difference between the Cohen's d effect sizes of LP vs. HP participants in both groups. More importantly, the LP in the experimental group had a more than double effect size than LP in the control group."

Fabic 2019 (2)¶

Fabic, G. V. F., Mitrovic, A., & Neshatian, K. (2019). Evaluation of Parsons Problems with Menu-Based Self-Explanation Prompts in a Mobile Python Tutor. International Journal of Artificial Intelligence in Education, 29, 507–535. https://doi.org/10.1007/s40593-019-00184-0

q3 · i? · causal · r2

Within the experimental group, LP students' normalized gains significantly exceeded HP students' on all questions (U = 66, p = .024), procedural (U = 68, p = .014), and easier-to-guess questions (U = 68.5, p = .011). LP post-test scores were not significantly different from HP students' post-test scores.

"LP students had a significantly higher normalized gain than HP students for all questions, procedural, and easier-to-guess questions."

Discussion¶

Related Claims¶

• High prior knowledge students scored significantly higher on SE prompts overall and on conceptual SE prompts, but not on procedural SE prompts — related
• Both experimental and control groups improved significantly on all questions and conceptual questions from pre- to post-test after using PyKinetic — related
• The experimental group spent significantly more time per problem and completed significantly fewer problems than the control group — related
• Both groups improved significantly on harder-to-guess questions, and the experimental group outperformed the control group on them at post-test — related