Across a geometry curriculum, overall percent correct stays fairly constant over time while percent correct on actions involving a single skill increases, as ACT-R's knowledge-component claim predicts¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2 · q2
Currency · newest study 2007 · not yet checked against its sources · not reviewed by a person · overdue for a recheck since 2010: tool and model accuracy changes quickly

Subclaims¶

q2 i? For 88 geometry students, percent correct remained fairly constant over time because rising task difficulty canceled out learning. → Ritter 2007 q2 i? When tracking only actions involving a single skill (area of a regular polygon), percent correct increased over actions, supporting learning at the knowledge-component level. → Ritter 2007 (2)

Evidence¶

Ritter 2007¶

Ritter, S., Anderson, J. R., Koedinger, K. R., & Corbett, A. (2007). Cognitive Tutor: Applied research in mathematics education. Psychonomic Bulletin & Review, 14(2), 249-255. https://doi.org/10.3758/bf03194060

q2 · i? · design · r2

A log-data analysis of 88 students using the Cognitive Tutor Geometry curriculum showed percent correct plotted over sequentially numbered actions remained fairly constant, which the authors attribute to increasing task difficulty canceling out learning.

"Figure 2 shows the percent correct, over time, for 88 students using the Cognitive Tutor Geometry curriculum in a school. The percentage correct remains fairly constant over time."

Ritter 2007 (2)¶

Ritter, S., Anderson, J. R., Koedinger, K. R., & Corbett, A. (2007). Cognitive Tutor: Applied research in mathematics education. Psychonomic Bulletin & Review, 14(2), 249-255. https://doi.org/10.3758/bf03194060

q2 · i? · design · r2

The same 88 students' actions restricted to a single skill (area of a regular polygon, one side horizontal) showed the increasing learning curve that ACT-R predicts for individual knowledge components; the article argues failure to see such a curve means the cognitive model misrepresents student knowledge.

"Figure 3 shows percent correct for the same group of students as in Figure 2, this time tracking only those student actions that the cognitive model considers to be relevant to a single skill (calculating the area of a regular polygon, in an orientation in which one side is horizontal)."

Discussion¶

Related Claims¶

• Learning-curve analysis of tutor logs revealed a flaw in the algebra cognitive model's handling of equations of the form ax = b when the coefficient is 1, and the tutor now provides targeted extra practice on that skill — related
• Note-level singing correctness rises with practice repetition in a negatively accelerated curve, gaining about 7 percentage points over 30 repetitions — related
• Simpler knowledge components are executed faster: correct application times rise from about 3-6 seconds for constant-constant KCs to 10-14 seconds for variable-variable KCs — related
• Trait-score trajectories over time are inconclusive as learning-curve evidence: one StudyChat trait increases while SWE-Chat traits stay flat — related